This web-exhibit should be cited as the following:
Kant, V., & Kashyap, A. (2021). Accidents in Mumbai 2014-2018. HFSS Studios. Unpublished Web-Exhibit.
Retrieved from
https://homepages.iitb.ac.in/~vivek.kant/indian-disasters/mumbai/
The data for this web-exhibit should be cited as the following:
Kant, V., Kushwaha, V., & Kashyap, A.
(2021). Accidents in Mumbai 2014-2018 [unpublished data set]. HFSS Studios. Retrieved from
https://homepages.iitb.ac.in/~vivek.kant/indian-disasters/mumbai/datasheet.xlsx
(The
dataset can be downloaded here)
"What is the life of an urban Indian worth?"
This may sound as an absurd question but it brings to fore an important challenge that is existing in the shadows of large cities in India. People often die from causes such as caving in of house roofs during rains, electrocution due to naked electric lines in slums, gas cylinders bursting in populated areas, among many other hazards that are interwoven in the life of poor urbanites. Living in dense cities in India is living in danger; this danger is a part of everyday existence. The malise is more insidious because it is accepted as part of everyday modern living. In other words, accidents that result in loss of lives are trivial and do not capture the public conscience.
We believe that these accidents are non-trivial and show the safety profile of the city. As a result, we have taken up the task of mapping these accidents, which typically are depicted on the backpages of newspapers. We aim to show that these accidents resulting in death and injuries have a story that needs to be told. Therefore, we started with this case study of charting out the accidents in Mumbai from the years 2014-2018 (half a decade), as reported in public newspapers. We focussed on public reporting of accidents that tells us the hidden loss of lives. While these may be depicted as trivial in public consumption of information, it also shows a safety profile of a city that needs sensitization and intervention towards the lives of urbanite Indians.
As a city of over 20 million people, Mumbai witnesses all kinds of disasters. These include natural as well as human made disasters, resulting in loss of lives on a daily basis. While these accidents have been well documented and received public recognition, there are a many accidents that do not make a dent in the public conscience. This case study highlights deaths and injuries that occur at commercial and residential sites in Mumbai between the years 2014 to 2018.
All the data for this case study has been compiled from daily reports from a single newspaper, The Times of India, Mumbai Edition (Click here to access the data source). It depicts accidents and injuries related to bizarre mundane life events which include a man and a 10-year old boy dying due to a wall collapse or more dramatic events such as ESIC Hospital fire in Kamgar. We have explicitly avoided traffic related accidents as these are already in the public awareness.
In this present case study, we have included visualisations that depict a range of insights. These include deaths across sectors such as commercial, factory, or residential, among others. The visualisations also include deaths and injuries throughout the years in relation to the days, months and quarters. The goal of our case study is data exposition for our readers, rather than providing an explanation about the causal basis for the injuries and accidents. To put it succinctly, we aim to reveal rather than explain.
We hope that this revelation will sensitise the public so that they demand a more detailed understanding and reduction in such accidents which can easily be prevented, saving many lives.
There are three legends primarily used to distinguish the dataset. The distinctions are made through time of the accident (shown in shades of orange for years and green for months), sector as classified by the newspaper (multi-coloured), and the reason for the accident (shown in shades of red for accidents involving fire and blue for other accidents).
Time labels have date, month and year wise distinctions; sector labels have Authority negligence, Commercial, Construction, Factory and Residential distinctions; reason labels have Building, Fire, Gas cylinder, Malfunction of machines and Negligence related accidents. These labels aid in recognising patterns and help assess the risks associated with the conditions in different times and areas. Labels for the sectors, as mentioned earlier, have been identified by the selected newspaper, labels for the reason have been identified as part of the research.
It is imperative to keep in mind that newspapers are aimed towrds a general audience and therefore may show some ambiguity in sector-wise labelling. These are reflected through variances in the the factors attributed in the reportage. For example, an incident involving a wall collapse in a market place can be labelled as both a commercial and construction accident. To ensure such miscrepencies do not creep into the dataset, the label for reason for the accident was created. The data was thoroughly cleaned and cross-checked with the original reports, and we adhered to the reasons as was reported in the newspapers.
The following graphs show different trends and insights related to deaths and injuries. These graphs are interactive in nature. We invite our readers to interact with them and try to understand the various challenges related to understanding the safety profile of Mumbai as regularly captured through newspapers.
Each nodal point can be interacted with individually. The date, number of casualties and any other relevant information are mentioned in the tool-tip. In addition, for further information, the view data button can be clicked, which will list all recorded facts and references related to the incident in question.
Our first visualisation below is a map of the accidents that occured between 2014 and 2018. The deaths and injuries have been marked, with the size of the circles represents the number of casualties involved, and the colour represents the time at which the incident took place.
In the maps below, the legends refer to the discrete number of deaths and injuries. For example, the clickable circle with the number 9 next to it demarcates all deaths and injuries in which a total of 9 people died or were injured. The map also reveals a greater number of dark circles, indicating that a larger percentage of accidents with casualties occurred towards the latter end of our dataset. This is more clearly illustrated in the Day-wise Analysis section of this case study.
The graph Deaths vs Injuries (Reason) compares the number of people dying to those getting injured. The reason we are comparing deaths to the injuries is to get an insight into the type of incidents that are high priority and are in immediate need of an intervention. The number of deaths is on the X-axis, and the number of injuries is on the Y-axis. In an ideal scenario, we would prefer all the data points to be located towards the bottom left of the graph. However, the graph shows that the deaths caused by Negligence* is disproportionately high. This indicates that this is an issue that is both dangerous and high priority. On the other hand, Gas cylinder related fires results in a higher percentage of injuries as compared to deaths, meaning there are more survivors from such accidents.
*The concept of negligence may be serving as an umbrella term for unique underlying issues that may not have been fully investigated and explicated yet. This is because newspaper reportage involves generic categories and a more common sense view of the matter as opposed to insights gained from scientifically rigorous analysis based on human factors.
The graph Deaths vs Injuries (Sector) is also a comparison between deaths and injuries, with the sector classified by the newspaper. Similar to the previous graph, an ideal scenario would have had all the circles towards the bottom left of the graph. However, we see that the Construction sector is to the far right and is much lower than all sectors but three. This is indicative of the higher risks involved in working in the construction sector, which has a higher proportion of deaths as compared to injuries, which is starkly different from the Commercial sector, which has a higher proportion of injuries as compared to deaths.
The graphs “Reason-wise breakdown of Sectors” and “Sector-wise breakdown of Reasons” shows the relationship between the sectors and the reasons attributed to the accidents (newspaper-identified).
The graph “Reason-wise breakdown of Sectors”, shows clearly that accidents related to buildings and fires result in the maximum number of casualties, both for deaths (60%) and injuries (70%). The problem resonates with the generic problems in the urban centres where construction related deaths and spread of fire causes havoc. The graph also indicates that necessary interventions are necessary to reduce deaths and injuries in Mumbai.
From “Sector-wise breakdown of Reasons”, the residential sectors show a marked influence due to the presence of gas-cylinder related accidents (number of injuries is more than the number of deaths in this sector). Improving gas-cylinder usage in urban centres like Mumbai is needed for reducing the number of casualties. In general, fires act as a major reason for casualties in the Commercial, Factory and Residential Sectors (47%), making it an area in immediate need for intervention to aid in the reduction of casualties.
The accidents, derived from the dataset, have been classified into five categories: Authority Negligence, Commercial, Construction, Factory and Residential. Those accidents that do not fall under any of these categories, are classified under Miscellaneous. The deaths per category have been shown to allow comparison between the sectors and the deaths across the years. The legend lists the years that the accidents were recorded. The casualties per year for each of the graphs can be seen by selecting the year from the legend.
As is seen in the graph (deaths vs sectors), there is no particular sector that shows the majority of the deaths, except construction. In this sector, the number of deaths are much higher than expected. While other sectors have recorded many deaths, it is important to recognise a sector in which there are known dangers still witnesses casualties in such large numbers; construction accidents remain the major cause of loss of lives. With Mumbai as a rapidly emerging city with ongoing urban renewal construction safety still remains a major priority to be addressed.
In this graph (injuries vs. sectors) we identify the injuries per sector to show that injuries outnumber the deaths. The commercial sector shows that the number of injuries is almost double to that of the number of deaths. However, in the construction sector, the proportion of deaths in relation to the injuries is quite high. In contrast, in the commercial sector, the proportion of deaths in relation to the injuries is much lower than that of construction, indicating that this might be a relatively safer sector. This also shows that in Mumbai as a metropolitan city, there may be a need for safety intervention in the construction sector in general.
To further comprehend the dangers of working the various sectors*, we mapped the casualties in the various sectors per month using treemaps. The area of the rectangle indicates the number of injuries for that particular month. The shade of blue indicates the number of deaths in the sector (the darker the shade, more are the deaths). As is seen clearly, construction (top right corner) is the only sector in which there is an association between the casualties from one month to another. The number of injuries in this sector is highest in the month of July, when the monsoon season starts. While the number of injuries decreases for the month of August, the number of deaths increases. This directly points to the lack of safe working conditions of the labourers in the construction sector during the expected seasonal rains.
*There are a few particular accidents that present a spike in the data resulting in certain sectors being overly represented due to this short duration between 2014 and 2018. For example, accidents such as the fire at ESIC Kamgar Hospital increased the number of injuries in the commercial sector in December, 2018 or the fire at a factory in Dombivli in May, 2016, accounted for more than a third of factory related injuries.
The following graph (Death vs Injuries by Months) helps illustrate the various trends across the newspaper categorised sectors clearly. The left axis represents the absolute number of deaths, the right axis represents the absolute number of injuries, and the bottom axis is labelled numerically with each number representing a month ordinally. The axis to the right is scaled up five times. This helps visualise the ratio of deaths to injuries in each month per sector. As the monsoon months begin, accidents due to Authority Negligence, the Construction and Residential sector note an increase in casualties. It is striking that in all three sectors, there exists a higher proportion of deaths when compared with the injuries. It is a lot more apparent in the construction sector, where there is a sharp spike in casualties in the month of July, and the ratio of deaths to injuries increases for the month of August.
The data can be broken down month-wise. This gives a histogram that indicates which of the months have a higher proclivity towards accidents and injuries. In the following graphs, the legend lists the sectors, which can be seen separately upon selecting them from the legend. The reason for listing month-wise is to observe developing monthly trends which result in casualties. For example, the monsoon months (June-September)result in a higher number of casualties (death and injuries)*.
*In the graph showing injuries per month, the trend is not apparent because there have been individual accidents that have contributed towards a higher number of injuries, for example, as discussed earlier, the Chemical factory fire in Dombivli in May, 2016 or ESIC Hospital fire in Kamgar in December, 2018.
The following graphs use years as the major separator. July is marked in green to showcase the onset of the monsoon season in India. The legend lists the months, which makes it possible to check for month-wise variation in casualties across the years. Many of the deaths in 2017 and 2018 were the result of some major accidents such as the Elphinstone Bridge collapse and ESIC Kamgar Hospital fire, which took many lives. Efforts should first be directed at minimising the loss of lives in such major accidents, by educating the public about accident etiquettes and incident awareness and ensuring public safety by enabling smooth functioning of public authorities responsible for prevention, mitigation and preparedness.
When using the individual days as separators, there appears to be a bit of a pattern, with many accidents occurring towards the end of the month in 2014-15 and 2017-18, while 2016 seems to be an outlier. Further analysis of more years may yield a more representative pattern. The days are listed in the legend, from 1 to 31, coloured in different shades of green. The darker the green, the later in the month the accident occured.
As was seen earlier, in "Death through the years", 2017 and 2018 are prominent in this visualisation. It is interesting to note the frequency of accidents in the second half of December for these two years.
A similar trend of accidents increasing towards the second half of the month can be observed in the graph "Injuries through the Years". Two major accidents are prominent, the Dombivli Factory fire in May 2016 and the ESIC Kamgar Hospital fire in December 2019 contributed to nearly 25% of the injuries in the past five years. This may be indicative of a browser trend that needs to be investigated further.
The following graphs display the 5-point summary of the dataset, in quartiles. The box-and-whiskers plots are primarily used to show the shape of a distribution, its central value and its variability. This variability is of central interest to us as it helps to illustrate how the different casualties in various sectors are distributed in relation to the anomalies. These anomalies can be easily identified by noting whether the dots representing the sectors lie within the expected error bars or beyond it.
For example, month-wise data for the variances and anomalies (Death and Injuries vs Sectors by Month) across the sectors is affected by seasonal changes. As discussed earlier, monsoons have a huge impact on the construction sector in such a grave manner that it goes beyond the accepted level of variances. This is indicative of the abnormally high casualty rates for this sector during the monsoon months. In the same graph, we also notice in September, Authority Negligence is beyond the top error bar. This is primarily due to the bridge collapse at Elphinstone Road in September of 2017. A similar phenomenon is visible in the Factory sector in May for injuries, with the Dombivli factory fire in May 2016 contributing as a prime factor for increase in casualties.
It is important to keep in mind that the dataset is gleaned from a national daily newspaper for a few years. Therefore, it may have missed out a few accidents. Nevertheless, the aim is to show that the safety profile of Mumbai involves a number of deaths that do not make headlines but are limited to backpages of a newspaper. The focus of this project is to bring to light accidents that do not make it to the headlines, for lack of sensationality, or for lack of sensitivity to the loss of human life. The hope is that this website would bring many trends and patterns, that are generally ignored, to the foreground, and help Indians citizens and policymakers to better prepare and reduce the loss of lives and livelihoods.
India's Disasters over the Decades