2

An anatomy of waste generation flows in construction projects using passive
bigger data

3

Jinying Xu1, Weisheng Lu1*, Meng Ye1, Chris Webster2, and Fan Xue1

1

4
5

6

7

1

Department of Real Estate and Construction, Faculty of Architecture, The University of Hong
Kong, Hong Kong, China
2

Faculty of Architecture, The University of Hong Kong, Hong Kong, China

*

Corresponding author
This is the peer-reviewed post-print version of the paper:
Xu, J., Lu, W., Ye, M., Webster, C., & Xue, F. (2020). An anatomy of waste
generation flows in construction projects using passive bigger data. Waste
Management, 106, 162-172. Doi: 10.1016/j.wasman.2020.03.024.
The final version of this paper is available at: https://doi.org/10.1016/j.wasman.2020.03.024.
The use of this file must follow the Creative Commons Attribution Non-Commercial No
Derivatives License, as required by Elsevier’s policy.

8

9
10
11
12
13
14
15
16
17
18
19
20
21
22

23
24

Abstract
Understanding waste generation flow is vital to any evidence-based effort by policy-makers or
practitioners to successfully manage construction project waste. Previous research has found
that accumulative waste generation in construction projects follows an S-curve, but improving
our understanding of waste generation requires its investigation at a higher level of granularity.
Such efforts, however, are often constrained by lack of quality “bigger” data, i.e. data that is
bigger than normal small data. This research aims to provide an anatomy of waste generation
flow in building projects by making use of a large set of data on waste generation in 19
demolition, 59 foundation, and 54 new building projects undertaken in Hong Kong between
2011 and 2019. We know that waste is generated in far from a steady stream as it is always
impacted by contingent factors. However, we do find that peaks of waste generation in
foundation projects appear when project duration is at 50~85%, and in new building projects
at 40~70% of total project time. Our research provides useful information for waste managers
in developing their waste management plans, arranging waste hauling logistics, and
benchmarking waste management performance.
Keywords: Construction waste management; waste generation flow; building projects; bigger
data

1

25

26
27
28
29
30
31
32
33
34
35
36
37
38

1. Introduction
Construction waste, also referred to as construction and demolition (C&D) waste, is the solid
waste resulting from construction, renovation, and demolition activities (HKEPD, 2018; Lu,
2019). In the US, it is estimated that 548 million tons of C&D debris, more than twice the
amount of municipal solid waste, were produced in 2015 (USEPA, 2018). According to WRAP
(Waste and Resources Action Programme) (2019), the construction industry is the UK’s largest
user of natural resources and generates 100 million tons of waste annually; over a third of the
country’s total waste. In China, it is estimated that C&D waste production will reach over 2.5
billion tons in 2020 (AECOM, 2018). In major economies worldwide with robust construction
sectors, 25~30% of solid waste landfilled comes from C&D activities (Hyder Consulting, 2011;
MoE, 2014; HKEPD, 2017). Given the adverse environmental impacts (e.g., greenhouse gas
emissions, leachate) of landfilling and its occupation of precious land that could otherwise be
used for urban development, it is clear that construction waste management is of vital
importance.

39

40
41
42
43
44
45
46
47
48
49
50
51
52
53

Measurement of construction waste generation is key to any effort to properly manage it (Lu
et al., 2016). Numerous studies have been conducted to quantify construction waste generation
at project, regional, and national level. In a comprehensive review, Wu et al. (2014) classify
these studies into six types: site visit, waste generation rate, life cycle analysis, classification
accumulation, variables modeling, and other. At regional or national level, quantification
studies can be used to help plan a city’s landfilling and other waste management facilities
(HKEB, 2013), estimate material stock in buildings (Kleemann et al., 2017), and approximate
waste recycling potential (Wang et al., 2019). However, most studies are conducted at project
level (De Guzmán Báez et al., 2012; Bakchan and Faust, 2019; Šomplák et al., 2019), where
C&D waste quantification has significant practical implications. It can provide critical
information for devising a waste management plan before construction has commenced, which
is becoming a standard practice in many economies (De Guzmán Báez et al., 2012; Šomplák
et al., 2019). It can also be used to extrapolate waste generation in a future project, with a view
to planning waste transportation logistics, or bidding for the project (Lu et al., 2016).

54

55
56
57
58
59
60

Nevertheless, quantification of C&D waste generation at project level is onerous and very often
constrained by lack of quality data. Prevailing construction practices do not require recordtaking of waste generation. Obtaining secondary data from a central source is not always
feasible, so researchers often have to collect firsthand data themselves. They can only do so
from a relatively small sample owing to the difficulties of conducting surveys of large-scale
building projects over a long period (Lu et al., 2018a). Difficulties also lie in the transient
2

61
62
63
64
65
66
67
68
69
70
71
72

nature of building projects (Demian and Walters, 2014). Unlike municipal solid waste
generated by households or a community in a steady stream, a building site ceases to generate
construction waste upon project completion. The project team, together with sporadic
construction waste data, is soon dispersed. An exceptional case is Lu et al. (2016), who
managed to collect a set of very good, “passive” data from 138 building projects in Hong Kong.
Using the data, they discovered that the accumulative waste generation as a project progresses
follows a sigmoidal or S-curve. While inspiring, this research has two shortcomings. First, it
does not cover foundation or demolition projects, which are non-negligible C&D waste
generators. Second, assessing waste generation flow (WGF) at a higher level of granularity and
at regular intervals, e.g., on a weekly or daily basis, is required for the in-depth understanding
needed to devise a comprehensive construction waste management plan. An anatomy of
construction WGF in various types of building projects is thus highly desired.

73

74
75
76
77
78
79
80
81
82
83
84
85

This research aims to offer an improved understanding of WGF in construction projects. It does
so by exploiting a large set of highly structured data on C&D waste generation from
construction projects in Hong Kong. The research is significant in that (a) improved
understanding of WGF is fundamental for construction waste management plans, e.g., onsite
and offsite treatment; and (b) the data analytics adopted will reveal useful information that
cannot be discovered with small data and encourage exploration of bigger data in construction
waste management research. The remainder of the paper is organized as follows. Section 2 is
a review of big and small data. Section 3 presents the materials and methods. Section 4
elaborates the analyses of the intra- and inter-groups of demolition, foundation, and building
(both residential and commercial building) projects and presents the results and findings.
Section 5 discusses the usage and significance of the results and findings and the
methodological implications of the study. Our conclusion is drawn in Section 6.

86

87

88
89
90
91
92
93
94
95

2. Literature review: Bigger data vs. small data
“Big data” is rapidly emerging in research disciplines including business, finance, management,
ecology, and medicine. It is a multifaceted concept subject to multiple definitions. Big data can
be understood by comparing it with “small data”; it is bigger. It is a collection of data so large
and complicated that it is difficult to process using traditional small data management tools.
Strictly complying with this definition, most so-called big data are just bigger data. While many
researchers stress the volume of big data, others, including Lu et al. (2018b), argue the
“relativeness” of big data. The volume of big data, be it gigabytes, or zettabytes, is a moving
target, depending on data generation capacity of the era (Everts, 2016); its strength is its ability

3

96
97

to present a fuller picture so as to have a closer claim of objective truth (Bilal et al., 2016; Lu
et al., 2018b).

98

99
100
101
102
103
104
105
106
107

The big data phenomenon can also be understood by differentiating passive and proactive data.
Chen et al. (2016) refer to the data of active solicitation (e.g., surveying or interviewing subjects)
as small data, and data generated unintentionally but with potential research uses as passive
(big) data. Currently, many big data sets are an unintended byproduct of business (Ekbia et al.,
2015), created for example by people traveling around with GIS-embedded devices,
communicating using smartphones, or purchasing through e-commerce (Bakchan and Faust,
2019). Since this passive big data is not carefully curated, it can better reflect an objective truth
as it happened (Lu et al. 2018b). Big enough, the data set can honestly record general business
done as well as randomness or outliers.

108

109
110
111
112
113

Traditional research on construction waste generation tends to rely on actively solicited,
carefully curated small data. Such data risks reflecting a small sample or a mere snapshot of
waste generation. This paper endeavors to collect and analyze bigger construction waste
generation data sets. Although not big enough to be called big data, they are definitely bigger
than small data and are expected to offer more confident or insightful research findings.

114

115

3. Materials and method

116

2.1 Data Collection

117
118
119
120
121
122
123
124
125
126
127
128

Since 2006, based on the polluter pays principle, the Hong Kong Environmental Protection
Department (HKEPD) has operated a Construction Waste Disposal Charging Scheme
(CWDCS). The CWDCS mandates that all construction waste, if not otherwise reused or
recycled, must be disposed of at government waste facilities (i.e. landfills, offsite sorting
facilities [OSFs], and public fill banks). The main contractor is charged a tipping fee of
HK$125 for every ton of inert waste it dumps in landfills; HK$100 per ton for mixed inert and
non-inert waste accepted by OSFs; and HK$27 per ton for inert waste accepted by public fills.
Prior to using these government waste disposal facilities, a main contractor
undertaking construction work under a contract valued HK$1 million or above is required to
open a billing account with the HKEPD solely for the contract, providing basic information
including contract name, contract sum, site address, type of construction work, etc. When the
construction waste is disposed of at the facilities, information on every load is recorded,
4

129
130
131
132

133
134
135

136
137

138
139

140
141
142
143

including vehicle number, time, weight of the vehicle upon entry and exit, and vehicle billing
account. The CWDCS thereby passively generates a big data set allowing investigation of
many aspects of construction waste management, including WGF. Appendix 1 illustrates the
structure of the big data set, which contains four types of databases as follows.
(1) The Project database contains all projects that have dumped waste in government facilities.
Recorded in this database are a total of 27,536 construction projects, with information on site
address, clients, project type, and other details.
(2) The Facility database contains all government construction waste management facilities,
e.g. landfills, OSFs, and public fills.
(3) The Vehicle database contains 9,863 waste hauling vehicles involved in construction waste
transport.
(4) The Waste Disposal database records every truckload of construction waste received at the
government waste management facilities. A total of 7,866,085 disposal records were generated
from all construction projects carried out during the eight-year period from 2011 to 2018, with
around 3,500 records being added every day.

144

145
146
147
148
149
150
151
152
153

Properly harnessed, this data set offers a better chance of investigating WGF because it
provides almost full coverage of waste generation from all sites in Hong Kong over the past
eight years. We obtained the data from the HKEPD’s general inquiry services and in recent
years developed a VBScript applet to automatically download the transaction records to a local
database for easier data access and storage. Receiving the passive data is a “trawling” exercise
and we developed “crawlers” to collect data from other sources, e.g., the Buildings
Department’s monthly project digest, as shown in the Building database in Appendix 2. We
can link our databases using indices. With this data on a longitudinal scale, we can track and
analyze WGF in different projects as they progress.

154

155

156
157

158
159

A set of qualifying projects was sourced from the pool according to the following criteria:
(1) building projects, either demolition, foundation, or new superstructure, as they form the
major physical structure of modern cities;
(2) sizeable projects of above-average contract sum for their kind, as they allow more regular
patterns than smaller counterparts; and

5

160
161

(3) must be completed and waste disposal activities recorded in the data set, i.e., started after
2011 and finished before 2019.

162

163
164
165

The projects are divided into three groups: demolition of old buildings or facilities, foundation
for new buildings, and new building construction (see Appendix 2). Comparison and
interpretation of projects’ WGF is conducted both intra- and inter-group.

166

167

3.2 Data preparation

168

3.2.1 Standardizing time

169
170
171
172
173
174
175

176

177
178
179
180
181
182

Since all projects differ in duration, the first step in data processing is to standardize time to
make them comparable. We count the time of a project from its first record of waste disposal.
For example, if the first record of Project A is week 27 of 2015, we calculate the project
duration according to this baseline. Supposing Project A ends at week 10 of 2017, then the
duration of Project A is 87 weeks (one year 52 weeks). Different projects may start from
different time points, which means their baselines are different. The time is then standardized
by percentage using formula (1) below:
%=

∗ 100%

(1)

Where ti is a time point of the project; T is the total duration of the project. Taking Project A
as an example, if T = 87 weeks, the standardized time of the second week is % = 2 87 ∗
100% = 2.3%. The standardization method is set on a weekly basis as a monthly basis is too
sparse. While the data available allows looking into WGF on a daily basis, having so many
days makes this “over-engineered”. It is too sparse to examine WGF on a monthly basis,
although the data also allows doing so.

183

184

185
186
187
188
189

3.2.2 Standardizing waste generation
As the weight of generated waste varies radically from one project to another depending on
size and other factors, the second step is to standardize waste generation. We treat the total
weight of waste generation of a project as 100%, and weekly generation as a percentage of the
total waste generation. We refer to this as waste generation ratio (WGR). The weekly waste
generation for a given project in a certain week is assessed by adding the weight of construction
6

190
191
192

193

194
195

waste truckloads (under the same billing account number) during that week. The weight of
weekly construction waste is also used to calculate accumulative WGR to a time point. The
WGR in a corresponding week of the project is calculated using formula (2):
%=

∗ 100%

(2)

where ri is the weekly WGR in week i, wi is the total waste generated in week i, and W is the
total waste generated in a project.

196

197

198

199
200

The accumulative WGR till week j is calculated using formula (3):
%= ∑

∗ 100%

(3)

where Apj is the accumulative waste generation ratio till week j, ri is the ratio of total waste
generation in week i, and i ranges from week 1 to week j.

201

202

203
204
205
206

3.2.3 Weekly and accumulative WGF curves
With the standardized time (Ti%) as the-x axis, weekly WGR (ri) as the-y axis, the weekly
WGF curve of a project can be drawn, as shown in Appendix 3(a). Replacing the weekly WGR
(ri) with accumulative WGR (Apj%), the accumulative WGF curve as the project progresses
can be portrayed, as shown in Appendix 3(b).

207

208

209
210
211
212
213

3.2.4 Scaling up
When the same approach is applied to all other projects, the weekly and accumulative WGF
curves of multiple projects can be sketched, as shown in Appendix 3(c) and 3(d), respectively.
Projects of the same group are arranged together. Consequently, a representative curve
epitomizing the general trend of projects of the same type is outlined for better interpretation
and comparison.

214

7

215

216
217
218
219
220
221

3.2.5 Inter-group curve
To compare individual and accumulative WGFs of the projects across their respective groups,
we need to develop a “representative” WGF curve for each group. A simple methodology is
adopted to produce the representative curves. At every 0.1% of the timeline, the waste
generation percentages of all projects within the same group are averaged to derive a point. By
plotting the points, the representative WGF curve of that group of projects is derived. We repeat
this methodology to derive the representative curves of all three groups.

222

223

224
225
226
227
228
229
230
231
232

3.3 Data analytical platform
Windows Microsoft Excel 2016 is capable of dealing with big data when the data is structured,
meaning that every record follows a standardized format. In Excel, our well-structured data
comprising 12,828 rows can be converted between different worksheets and calculated easily
using simple formulas embedded in the software, and figure plotting and manipulation is also
very convenient. Therefore, Excel 2016 is used to handle the data set extracted using the
VBScript applet from the raw data sets. The data is disaggregated first based on project group
and then project ID. Data preparation and analysis is conducted at individual project level and
then repeated for every project. For comparison purposes, all figures for each group of project
are integrated into one figure for group analysis.

233

234

235
236
237

4. Data analyses, results, and findings
In this section, weekly and accumulative WGFs of demolition, foundation, and new building
projects are analyzed within their respective groups. An inter-group analysis is also conducted
to compare the similarities and differences between different types of projects.

238

239

240
241
242
243
244
245
246

4.1 Demolition projects
Waste generation data for 19 demolition projects was collected for analysis. The shapes of the
WGF curves of these projects vary with some obvious outliers (see Fig. 1). There are three
WGFs (D1, D2, and D3) with only three points changing dramatically. Upon checking the
project database, we learned that they are projects with a relatively small contract sum, i.e., a
few million HKD compared to tens of millions for other projects, with projects D1 and D2 for
re-roofing and demolition of market stalls in two public housing projects, respectively, and D3
demolition and hoarding works in a private demolition project. Most of the projects have
8

247
248
249
250

similar WGRs, generating around 10% of the total waste every 5% of the project time. Within
each demolition project, it is hard to detect any stages that generate more waste than others. In
other words, waste generation in demolition projects is a steady stream that is more or less
evenly distributed from start to end.

251

252

253

Fig 1. Weekly waste generation flows of 19 demolition projects

254

255
256
257
258
259
260
261
262
263
264

The accumulative WGFs of the demolition projects are shown in Fig 2. These accumulative
WGFs follow an S-curve shape, starting slowly at the start, accelerating in the middle, and then
tailing off when a project nears its end (PMI, 2013). However, the shapes of the S-curves are
different from each other. Some of them are steep while others very smooth. There is one
special case (D1 in Fig. 2) where nearly 80% of the waste was disposed of in the first week.
This is because it is a small demolition project producing only 54.43 tons of waste in total with
43.01 tons generated in the first week lasting for only 12 weeks. Another outlier (D4 in Fig. 2)
is the demolition project for a standard 24-classroom, 6-floor primary school in a housing estate
and erection of associated hoarding and a covered walkway for residential use. At 60% project
duration, it had produced less than 2% of its total waste.

265

9

266

267

Fig 2. The S-curves of accumulative waste generation of 19 demolition projects

268

269
270
271
272
273
274
275
276
277

Further analyses were conducted to examine accumulative waste generation at critical time
points. At 25% project duration, seven out of the 19 projects had generated more than 25% of
total waste and four more than 50%. At 50% project duration, 11 projects had produced more
than 50% of total waste, four of which had already produced nearly 90% of their total waste.
At 75% project duration, 13 projects had generated more than 75% of their total waste. These
figures clearly show that different demolition projects produce waste at different speeds.
However, for most, waste generation in the middle stage is faster than in early and late stages.
Monitoring this critical time point can help predict and control waste generation for both waste
management facilities and contractors.

278

279

280
281
282
283
284

4.2 Foundation projects
Waste generation data for 59 foundation projects was extracted from the data set. Their WGFs
are shown in Fig 3. Some patterns are significant. The WGRs of most projects are under or
around 10%, mainly due to the long duration of foundation projects. On average, the foundation
projects took around 78 weeks to complete (see Appendix 1), so weekly waste generation is
lower than demolition counterparts which averagely last 40 weeks. Moderate in the early stages,
10

285
286
287
288
289
290
291
292
293
294
295
296

WGRs of foundation projects peak between 50% and 85%. Originally, it was thought that most
waste is produced during the early stages when evacuation and piling occurs. However, the
WGFs in Fig. 3 illustrate that the middle to later stages, when backfilling and demoulding take
place, generate the most waste. A possible explanation is that the evacuated soil can be used
for backfilling. Hence, in projects where a storage area exists, the soil will be kept until
backfilling is complete. Moreover, demoulding will inevitably create piles of mouldboard
waste. Where the site area is limited, waste has to be disposed of promptly. There are some
outliers in the WGF curves. A particular case is project F1 as highlighted in Fig. 3, where the
waste percentage remains at a higher level (above 20%) during 75~80% of total project
duration. More than 74% of F1’s waste was disposed of in three continuous weeks, a possible
explanation being that intensive evacuation works can generate a large amount of waste in a
short period of time.

297

298

299

Fig 3. Weekly waste generation flows of 59 foundation projects
11

300

301
302
303
304
305
306
307
308
309
310
311
312

The accumulative WGFs of foundation projects are shown in Fig 4. The shapes of the S-curves
do not overlap, proving that no two projects are the same and even though they have been
standardized. Some patterns can be easily detected. For most projects, the curves grow slowly
in the first 30% and final 20% of total time but climb quickly in the middle stages. An outlier
in Fig. 4 is F2, a massive piling foundation project with a contract sum of HK$279 million
which had generated 92% of its total waste at 18% of project time. Another notable outlier, F3,
with a contract sum of HK$216 million only generated 1% of waste in the first 55% of project
time. Another project, F1, only generated 0.42% of waste within the first 44.4% of project time.
In some other special projects the S-curves have a flat line segment. One foundation project for
public rental housing development had almost no waste increase from 40% to 67% of the time,
after which its curve takes nearly a steep shape. This can be explained by the fact that in some
foundation projects, preparation works take months before evacuation work takes place.

313

314

315

Fig. 4. The S-curves of accumulative waste generation of 59 foundation projects

316

12

317
318
319
320
321
322

Further analyses were conducted to examine accumulative WGFs at some critical time points.
As shown in Fig. 4, at 25% of the time, only 9 out of 59 projects had produced more than 25%
of their total waste; at 50% of the time, 15 projects had produced more than 50% of waste; and
at 75% of the time, 13 projects had generated more than 75% of waste. For 54 (more than
91.5%) projects, the WGR is over 90% at 90% of the time. For the majority, waste generation
is relatively slow for the first half of the project.

323

324

325
326
327
328
329
330
331
332
333
334
335

4.3 New building projects
Waste generation data of 54 selected new building projects was obtained. Their WGFs are
shown in Fig. 5, where some patterns are noticeable. Apart from several outliers, the
distribution of the ratios is more converged than for demolition and foundation projects. The
highest WGRs of most new building projects, at less than 5%, are smaller than those of
foundation projects. This is primarily due to the difference in completion time between new
building projects (average 199.2 weeks) and foundation projects (average 77.9 weeks), as
shown in Appendix 1. Unlike foundation projects with relatively concentrated peaks between
50% and 85% of project duration, these peaks appear in new building projects mostly between
40% and 70%. The WGFs of new building projects also have an early peak at 20% of the
project time because site leveling works at the start of new building projects generate a large
amount of waste.

336

337
338
339
340
341
342
343
344
345
346
347
348
349

The four observable outliers are: B1 (very early peak), B2 (peak value 36.75%), B3 (shorter
time with dramatically changing value), and B4 (two high peaks). B1 is a residential
development project comprising 17 blocks of 6-storey, low-rise buildings on Lantau Island.
Simple construction requirements and a relatively capacious site made construction waste
storage possible at the start of the project, leading to disposal of about 22% of total waste
(mainly from site levelling) in one week at around 6% of progress. B2 is an 88-week residential
building project with a contract sum of HK$239.8 million. At week 85, 36.75% of its total
waste was disposed of in one week, much different than others that dispose wastes consistently.
B3 is also a residential building project with a contract sum of HK$368.9 million and a duration
of 52 weeks. No waste was disposed of from weeks 2 to 13, but with a dramatical changes of
waste disposal amount at middle stages. B4, a 63-week project with a contract sum of HK$251
million is located in Yuen Long, a relatively rural and expansive area allowing some waste
storage on site for disposal once for a long interval, leading to the two high peaks at the WGF.

350

13

351

352

Fig. 5. Weekly waste generation flows of 54 new building projects

353

354
355
356
357
358
359
360
361
362

The accumulative WGFs of new building projects are shown in Fig. 6. Some general patterns
can be observed. Although the shapes of the S-curves differ, most are parallel at the middle
stages. This means the accumulative WGF patterns are similar. This pattern implies that Hong
Kong contractors are quite adept at superstructure building projects, which are usually less
impacted than other groups of projects by contingent factors such as design change, road
congestion, or extreme weather. In the first 30% and final 30% of project time, the WGRs of
most projects are moderate. There are 11 projects (nearly 20%) that produced more than 60%
of total waste in the first 30% of the time and only 5 (less than 10%) that had produced less
than 70% of total waste at 70% of project duration.

363

14

364

365

Fig. 6. The S-curves of accumulative waste generation of 54 new building projects

366

367
368
369
370
371
372
373
374
375
376
377
378
379
380

From the horizontal perspective, at 25% of project time, 16 out of 54 projects (nearly one-third)
had produced more than 25% of their total waste. At 50% of project time, 37 projects (more
than half) had produced more than 50% of their waste. At 90% of time, 53 projects (more than
98%) had generated more than 90% of their total waste. Most projects generated waste very
quickly in the period 25% to 75% of project time. Unlike our casual remarks that waste
generation in a high-rise superstructure building project would be a steady stream, WGFs
actually concentrate in the middle 40~70% stage. An outlier in Fig. 6 is B5, which only took
34.5% of the time to produce 99.5% of its waste. B5 is a massive public rental housing
development project with a HK$1.75 billion contract sum and lasting four years. It adopted
prefabricated components and precast elements such as fabric reinforcement, semi-precast slab,
precast façade and staircases, and also volumetric precast kitchen and bathroom, the
implementation of which reduced wastage when constructing standard floors. Another outlier
is the earlier-mentioned B4, where only 6% of waste had been disposed of at 55% of
completion time because the wide site allowed for waste storage.

381

15

382

383
384

4.4 Inter-group analyses
This section compares individual and accumulative WGFs of the projects across their
respective groups, as shown in Figs. 7 and 8.

385

386

387

Fig. 7. Comparison of average individual WGFs of three types of projects

388

389

390
391
392
393
394

395
396
397
398
399

400
401
402

Several patterns are observed in Fig. 7, as follows:
(1) Generally, the three types of projects present different patterns of WGFs. The WGFs of
foundation and new building projects display a reversed U shape as the projects progress while
the WGF of demolition projects fluctuate randomly. The peak of the reversed U in foundation
projects happens at around 70% of progress while in new building projects it happens earlier
near 50%. In contrast, demolition projects generate a large quantity of waste at their final stage.
(2) Demolition projects’ WGF fluctuates the most among the three types of projects. These
projects generate waste more randomly at different stages of the project. The individual WGFs
shown in Fig. 1 reveal no consistent pattern. There is no “standard” construction waste
management method for demolition projects; rather, it depends on the nature of the project, the
site, and the demolition methods adopted.
(3) Foundation projects generate most construction waste in their middle and later stages,
especially during 55~85% of project time. This can be explained by excavation, backfill, and
residual of waste materials being typical waste generation processes in foundation construction.
16

403
404
405

406
407
408
409
410

Based on this pattern, contractors and waste management facilities should reserve enough
capacity to store, transport, and dispose of waste that is quickly generated at a later stage of
these projects.
(4) New building projects generate large quantities of waste at the early stage, i.e. 5~10% of
project time. Most construction waste is generated in the middle stage, at 40~70% of project
time. Afterwards, the WGR declines until the final stage of the project. In this regard,
contractors need to pay attention to the WGFs and arrange their waste management plans
accordingly.

411

412
413
414
415

It will be informative to compare the cumulative WGF curves of the projects across their
respective groups. By repeating the methodology as described above for individual WGF
curves, the representative accumulative WGF curves of the three groups are derived, as shown
in Fig. 8.

416

417

418

Fig. 8. Comparison of average accumulative WGFs of the three groups of projects

419

17

420

421
422
423
424
425
426

As illustrated in Fig. 8, all curves are S-curves but their shapes differ. Patterns observed are:
(1) Generally, the curve of new building projects is higher than that of foundation projects
(except at the final stage). The gap between the new building and foundation project curves is
small at the beginning and increases and then narrows until they converge at around 90% of
progress. New building projects produce a bigger share of waste than foundation projects. New
building projects generate more waste at the early stage (the first half of progress) while
foundation projects produce more waste at the later stage (the last half of progress).

427

428
429
430
431
432
433
434

During the first 44% of project progress, demolition projects generate waste faster than new
building projects but the generation ratios gradually decrease. At 56% to 87% of project time,
new building projects accumulate 10% more waste than demolition projects. Demolition
projects also produce waste faster than foundation projects in the first 72% of project time. The
gap remains larger than 10% during the period 22% to 55% of project progress. For the last
28% of the time, foundation projects generate more waste than demolition projects. The gap is
the widest at 86% of project time.

435

436
437
438
439
440
441
442
443
444

(2) The difference between different groups of projects changes over time. In the first 25% of
time, the accumulated waste percentage for demolition, foundation, and new building projects
is 26.8%, 15%, and 24.5% respectively. At 50% of progress, new building projects have
generated nearly 60% of their total construction waste, demolition projects 55%, and
foundation projects 42%. At 75% of project time, the accumulated waste percentage for
demolition, foundation, and new building projects is 79%, 82%, and 92%, respectively.
Demolition projects produce waste quickly in the early stages; new building projects create
more waste in the early-middle stages; and foundation projects generate most of their waste at
the middle and later stages.

445

446

5. Discussion

447

5.1 Improved understanding of construction waste generation

448
449
450
451
452

Findings of this research well resonate with previous studies. Particularly, the findings of this
research echo Lu et al. (2016) in finding that the accumulative WGFs in building construction
indeed follow an S-shape curve. A new finding is that the WGFs in demolition and foundation
projects also follow an S-shape curve. Furthermore, the S-curves, compared either intra- or
inter-group, show some interesting patterns previously unknown. From an inter-group
18

453
454
455
456
457
458
459
460
461
462
463
464
465

perspective, new building and foundation projects have generated around 50% of their total
waste when they approach 50% of the project schedule, while demolition projects as a whole
have generated only 40% of the total waste at the same point in time. Knowing that construction
waste generation in a project is a non-linear process with a steady stream (De Guzmán Báez et
al., 2012), a construction project is often impacted by all kinds of factors, including design
change, material shortage, accidents, road congestion, and extreme weather, and the WGFs
reflect this. However, the overall patterns of different groups of projects are intriguing, which
might be explained by the general nature of the projects. Probing the accumulative WGFs of
each group, it is noticed that their accumulative waste generation varies significantly from one
project to another, with some really distributing as outliers. Some may have generated more
than 70% of total waste halfway through the project, while others may have generated only
30% at the same point. To understand the factors (e.g., project characteristics) that cause the
difference would be interesting.

466

467
468
469
470
471
472
473
474
475
476
477
478
479
480
481

This research also looks at individual WGFs at regular intervals. The results show a difference
in WGFs across the three groups of construction project. WGFs of demolition projects fluctuate
with no outlier peaks; foundation projects have one major peak between 50~85% of progress;
new building projects have a minor peak at 20% and a major peak during 40~70% of the project
progress. Different demolition projects may generate construction waste at different speeds
throughout their progress. This is dependent on demolition methods which are, in turn,
determined by site condition and volume of demolished buildings or structures. For foundation
projects, since the foundation types of residential buildings are not diverse, WGF remains
moderate during evacuation and piling works and peaks during backfilling and demolding in
the middle to later part of the project. This finding contradicts the claim in Yu et al. (2013) that
the earthworks stage in a foundation project produces most of the waste. In new building
projects, site leveling at the initial stage creates tons of waste soil, causing a minor peak in
WGF. Afterwards, waste is created slowly. Although the individual and accumulative WGFs
of a project are mutually convertible, the individual WGFs as discovered in this paper can better
improve our understanding than previous similar studies using linear or sigmoidal models.

482

483

484
485
486
487
488

5.2 Applications of the results
Our research findings contain some latent knowledge that can inform practical waste
management applications. For example, the WGF and accumulative WGF curves can help
predict the generation of waste at different stages of different types of projects, enabling
planning of materials, workforce, and vehicles in advance. For top-down demolition projects
using manual methods or machines, an opening will usually be made to allow machines to be
19

489
490
491
492
493
494

lifted to the top floor (see Appendix 4 for details). This opening can then also be used to
transport demolition waste from upper to lower floors, which can be used as a buffer for storing
the waste. Disposing of the waste thus becomes less of a pressing issue and more attention can
be paid to other issues, e.g., safety, noise, and vibration hazards frequently reported in
demolition projects. Intensive waste hauling services can be planned when demolition projects
approach 90% of their progress.

495

496
497
498
499
500
501
502
503
504
505

Foundation projects, especially deep foundations for high-rise buildings widely seen in Hong
Kong, need careful planning of piling, excavation, anchoring, and backfilling in what is
normally a very much confined site area (see Appendix 5). Space for temporarily storing waste
must compete with space allocated for other activities such as placing materials or installing
machinery. The WGF curves of foundation projects as shown in Fig. 4 are distributed widely.
These projects are often impacted by contingent factors, such as unexpected geotechnical
conditions, water leakage, and uneven settlement (Canonico & Söderlund, 2010). On the other
hand, reuse of excavated soil is desired to save construction costs. The WGFs are not only to
provide a converged pattern for waste management but also do warn of the potential impacts
of contingent factors that should be given careful consideration.

506

507
508
509
510
511
512
513
514

For new buildings projects, what is against our casual remarks is that waste generation amount
is not as much as their demolition or foundation counterparts. There is a minor peak at 20% of
project duration, mainly owing to site formation and non-standard floor construction. After the
project enters the standard floor construction, some of the built-up floors can be used for
temporary storage. The waste generated in upper floors can be moved to lower floors or ground
floor storage directly through a debris chute (see Appendix 6). The curves shown in Fig. 5
demonstrate that major waste generation peaks at 40~70% of the project progress. Therefore,
careful planning of onsite storage and waste hauling services should be targeted in this period.

515

516
517
518
519
520
521

In prevailing practice, a waste management plan is formalized as a project document before a
project commences. However, the knowledge for making such plans often resides in the mind
of experienced managers and is not necessarily accessible to others. It is not uncommon for
managers to alter old waste management plan for a new project. The improved understanding
of waste generation flows derived from our data analytics can help to make more effective
plans to manage this waste.

522

20

523
524
525
526
527
528
529
530

Our results are transferrable to projects both in and outside Hong Kong. The large volume of
the sample studied could reduce the unique characteristics of individual projects and keep the
common waste generation pattern of all projects of the same kind, while the same types of
construction projects have similar practical procedures and methods. However, to validate the
findings, further studies using the same research methods are desired to compare the results of
this paper with practices from elsewhere. Specifically, for example, future research may
investigate how different materials and technologies across different geographic contexts
impact the waste generation flows and patterns.

531

532

533
534
535
536
537
538

5.3 Applications of bigger data
Passive bigger data is useful for achieving an anatomy of the WGFs of completed construction
projects. In comparison with small data, bigger data can mitigate the unique features of
different projects, alleviate the impacts of outliers and reveal regularities. Without passive
bigger data, immense effort would have to be spent on collecting firsthand data for just a single
curve, let alone the 132 curves of this paper, and an understanding of WGFs in different types
of projects would be extremely difficult to achieve.

539

540
541
542
543
544
545
546
547
548
549

Looking at the curves drawn, the outliers are even more informative than the converged curves.
For example, we know that project B5 in Fig. 6 adopted a series of low-waste construction
technologies. However, without the bigger data to paint a fuller picture, B5 would not stand
out in terms of WGF. However, bigger data does not complete the picture. For example, the
bigger data showed that project B2 in Fig. 5 disposed of 1812.7t of waste in a week. Google
Maps reveals that the project is located on a compact site. Without project knowledge, it is not
possible to interpret this huge volume of waste disposal in such a short time period. To fully
harness the power of big data analytics, one needs to combine bigger data with “thick data”:
meaningful qualitative data on behavior and its underlying motivations (Rasmussen and
Hansen, 2015).

550

551
552
553
554
555

Bigger data use has a potential caveat. It is assumed that the waste disposal represents how
waste is generated when a trade is made. This hypothesis is established under the special
conditions of Hong Kong, where construction sites have little to no space for waste storage onsite, so waste generated has to be disposed of as soon as possible. With this potential caveat,
the bigger data and its analytics demonstrate a compelling technique to probe into construction

21

556
557

waste generation, with a view to providing better decision-making information for waste
management.

558

559

560
561
562
563
564
565
566
567
568

5.4 Limitations
The first limitation of this study is the context of the analyzed projects. Since all the data are
collected from Hong Kong projects, the findings best report waste generation patterns of Hong
Kong projects. Although the study can work as a reference and guidance for similar projects in
other contexts, validation is needed for further confirmation. The second limitation is that no
two projects are the same. Although we tried to select comparable projects, they are still
different to a greater or lesser extent. Meanwhile, the big sample makes it difficult to investigate
the detailed characteristics of individual projects. These two weaknesses limit the findings of
the study. Further research will be done to overcome them by triangulating bigger data with
small thick data.

569

570

571
572
573
574
575
576
577
578
579
580
581
582
583

6. Conclusion
Using a set of passive bigger data on waste disposal in Hong Kong, the authors provide an
anatomy of waste generation in three types of construction project: demolition, foundation, and
new building projects. We find that the WGFs of each group show quite different patterns. The
WGFs of demolition projects fluctuate with no outlier peaks; foundation projects have one
major peak at 50~ 85% of project progress; and new building projects have a major peak during
40~70% of progress and a minor peak at the first 20%. Demolition projects may generate
different but largely steady streams of waste throughout their progress. WGFs in foundation
and new building projects largely follow a reversed U shape as the projects progress.
Unexpectedly, WGFs in foundation projects remain moderate during piling and evacuation
works, but reach their peak during the middle to the later stage of the projects when backfilling
and demolding takes place. In new building projects, site leveling at the initial stage will cause
a minor peak in WGFs. Afterwards, waste generation reaches a major peak during 40~70% of
the total project schedule.

584

585
586
587
588

This research echoes previous studies to confirm that accumulative waste generation in
construction projects follows an S-curve. We find that the shapes and trends of the S-curves of
the three types of project have considerable differences. New building and foundation projects
have more stretched curves than demolition projects. The accumulative WGFs in foundation
22

589
590
591
592

and new building projects have fewer radical changes compared to demolition projects. In most
foundation projects, half of the total waste is created quickly in the middle and later stages (at
50~75% of project progress). For most new building projects, 25~90% of total waste is
generated in less than one-quarter of the project time, concentrating at 40~65% of project time.

593

594
595
596
597
598

This anatomy offers an improved understanding of WGFs in demolition, foundation and new
building projects. It provides insightful information that can be used for, e.g., extrapolating
waste generation, developing waste management plans, and arranging waste hauling logistics.
Some outlying cases or anomalies are vividly visualized against the rest of projects for
benchmarking performance or indicating areas for further investigation.

599

600
601
602
603
604
605
606
607
608
609
610
611
612

We use typical big data analytics processes, such as data collection, cleansing, processing, and
visualization. The paper does not involve cumbersome statistical methods but relies on
visualization, offering superior explanatory power that cannot be derived using small data.
Bigger data can reveal regularity amongst a multitude of cases by alleviating the impacts of
outliers to show the general patterns, and helps us investigate the collective WGFs of different
types of projects. Outliers in bigger data are not often considered but are actually informative.
However, without bigger data to paint a full picture, these outliers would not be standing out
to notice without the comparison with others. Bigger data is not a silver bullet. It can allow
some useful information to surface. To understand the causes behind the various patterns, one
needs to combine bigger data with small, quality, informative “thick data” by further
investigating the subject matter. Specifically, for example, future research may investigate
what factors underlie the differences in waste generation curve shapes, and how different
materials and technologies across different geographic contexts impact these shapes.

613

614

615
616
617
618
619

Acknowledgement
This research is supported by the General Research Fund (GRF) (Project No.: 17201917) from
the Hong Kong Research Grants Council (RGC), and Public Policy Research (PPR) (Project
No.: 2018.A8.078.18D) and Strategic PPR (Project Number: S2018.A8.010) Funding Schemes
from the Policy Innovation and Co-ordination Office of the Government of the Hong Kong
Special Administrative Region.

620

23

621

622
623
624

625
626

627
628
629

630
631
632

633
634
635

636
637
638

639
640

References
AECOM. 2018. People’s Republic of China: Construction and Demolition Waste Management
and
Recycling.
https://www.adb.org/sites/default/files/projectdocuments/48105/48105-001-tacr-en.pdf (Accessed 16 January 2020).
Bakchan, A., & Faust, K. M. 2019. Construction waste generation estimates of institutional
building projects: Leveraging waste hauling tickets. Waste Management. 87, 301–312.
Bilal M., Oyedele L.O., Qadir J., Munir K., Ajayi S.O., Akinade O.O., et al. 2016. Big Data in
the construction industry: A review of present status, opportunities, and future trends.
Advanced Engineering Informatics. 30(3), 500–521.
Canonico, P., & Söderlund, J. (2010). Getting control of multi-project organizations:
Combining contingent control mechanisms. International Journal of Project
Management, 28(8), 796-806.
Chen C., Ma J., Susilo Y., Liu Y., Wang M. 2016. The promises of big data and small data for
travel behavior (aka human mobility) analysis. Transportation Research Part C:
Emerging Technologies. 68, 285–299.
De Guzmán Báez A., Sáez P.V., Del Río Merino M., Navarro J.G. 2012. Methodology for
quantification of waste generated in Spanish railway construction works. Waste
Management. 32(5), 920–4.
Demian P., Walters D. 2014. The advantages of information management through building
information modelling. Construction Management and Economics. 32(12), 1153–1165.

643

Ekbia H., Mattioli M., Kouper I., Arave G., Ghazinejad A., Bowman T., et al. 2015. Big data,
bigger dilemmas: A critical review. Journal of the Association for Information Science
and Technology. 66(8), 1523–1545.

644

Everts S. 2016. Information overload. Distillations. 2(2), 26–33.

641
642

645
646

647
648

HKEB. 2013. Hong Kong Blueprint for Sustainable Use of Resources 2013–2022.
https://www.enb.gov.hk/en/files/WastePlan-E.pdf (Accessed 3 June 2019).
HKEPD. 2017. Construction Waste Disposal Charging Scheme.
https://www.epd.gov.hk/epd/misc/cdm/scheme.htm (Accessed 1 June 2019).

24

649
650
651

652
653
654

655
656

657
658

659
660
661

662
663
664

HKEPD. 2018. Monitoring of Solid Waste in Hong Kong — Waste Statistics for 2017.
https://www.wastereduction.gov.hk/sites/default/files/msw2017.pdf (Accessed 1 June
2019).
Hyder Consulting. 2011. Construction and Demolition Waste Status Report — Management
of Construction and Demolition Waste in Australia. https://bit.ly/2nZfFWm (Accessed
3 June 2019).
Kleemann F., Lederer J., Rechberger H., Fellner J. 2017. GIS‐based analysis of Vienna’s
material stock in buildings. Journal of Industrial Ecology. 21(2), 368–80.
Lu W., Peng Y., Chen X., Skitmore M., Zhang X. 2016. The S-curve for forecasting waste
generation in construction projects. Waste Management. 56, 23–34.
Lu W., Webster C., Peng Y., Chen X., Chen K. 2018a. Big data in construction waste
management: Prospects and challenges. Detritus. https://doi.org/10.31025/26114135/2018.13737
Lu W., Chen X., Peng Y., Liu X. 2018b. The effects of green building on construction waste
minimization: Triangulating ‘big data’ with ‘thick data’. Waste
Management. 79,
142–52.

666

Lu W. 2019. Big data analytics to identify illegal construction waste dumping: A Hong Kong
study. Resources, Conservation and Recycling. 141, 264–72.

667

MoE.

665

668

669
670

671
672

673
674
675

676
677

2014. History and Current State of Waste Management in Japan.
https://www.env.go.jp/en/recycle/smcs/attach/hcswm.pdf (Accessed 10 August 2019).

PMI. 2013. A guide to the project management body of knowledge (PMBOK guide). In Project
Management Institute (Vol. 5).
Rasmussen M.B., Hansen A.W. 2015. Big Data is only half the data marketers need. Harvard
Business Review, 16.
Šomplák R., Kůdela J., Smejkalová V., Nevrlý V., Pavlas M., Hrabec D. 2019. Pricing and
advertising strategies in conceptual waste management planning. Journal of Cleaner
Production. 19, 118068.
USEPA. 2018. Advancing Sustainable Materials Management: 2015 Fact Sheet.
https://bit.ly/2LGgoXj (Accessed 1 June 2019).

25

678
679
680

681
682

683
684

685
686
687

Wang J., Wu H,. Tam V.W., Zuo J. 2019. Considering life-cycle environmental impacts and
society's willingness for optimizing construction and demolition waste management fee:
An empirical study of China. Journal of Cleaner Production. 206, 1004–14.
WRAP. 2019. Guidance for small and medium sized contractors: Reducing your construction
waste. https://bit.ly/36wYTxZ (Accessed 13 August 2019).
Wu Z., Yu A.T.W., Shen L., Liu G. 2014. Quantifying construction and demolition waste: An
analytical review. Waste Management. 34(9), 1683–92.
Yu A.T.W., Poon C.S., Wong A., Yip R., Jaillon L. 2013. Impact of construction waste disposal
charging scheme on work practices at construction sites in Hong Kong. Waste
Management 33(1), 138–46.

688

26