Difference between revisions of "IS428 2016-17 Term1 Assign2 Teo Hui Min"
(53 intermediate revisions by the same user not shown) | |||
Line 1: | Line 1: | ||
=Theme & Questions= | =Theme & Questions= | ||
− | The | + | The dataset used for this assignment is the Workplace Injury in 2014. The focus of this assignment is to look for correlation between the injuries and the other variables, and looking into the injuries of 3 industries (Manufacturing, Construction & Others) from various viewpoint. Through the data discovery, it aims to answer the following questions: <br> |
− | 1. | + | 1. Is there a correlation between the experience of employees and workplace injury? Does age plays a part too?<br> |
2. What are the most common type of accident and injuries in each industry and how serious are the injuries?<br> | 2. What are the most common type of accident and injuries in each industry and how serious are the injuries?<br> | ||
− | 3. Which are the top 3 companies | + | 3. Which are the top 3 companies with the highest number of victims? <br><br> |
=Data Exploration= | =Data Exploration= | ||
+ | ==Transformed variable - No. of MC days== | ||
[[File:hm2_2.PNG | 1000px]] <br> | [[File:hm2_2.PNG | 1000px]] <br> | ||
+ | [[File:hm2_1.PNG | 400px]] <br><br> | ||
During the data exploration, it was discovered that some of the variables are skewed. One of the variables is ‘No. of MC days’, hence it was transformed by taking ‘LOG’. <br><br> | During the data exploration, it was discovered that some of the variables are skewed. One of the variables is ‘No. of MC days’, hence it was transformed by taking ‘LOG’. <br><br> | ||
− | |||
− | |||
[[File:hm2_3.PNG | 100px]] <br><br> | [[File:hm2_3.PNG | 100px]] <br><br> | ||
− | According to the MOM website, the maximum paid hospitalisation leave is only 60 days. Thus there may be a case of human error in the data entry. However the number of MC that employees are entitled varies, thus these anomalies are not being removed from the data. A filter will be implemented to allow users to have the autonomy to decide the range of MC days. <br><br> | + | However the range for this variable is 1 to 3600 days which actually does not make sense. According to the MOM website, the maximum paid hospitalisation leave is only 60 days. Thus there may be a case of human error in the data entry. However the number of MC that employees are entitled varies, thus these anomalies are not being removed from the data. A filter will be implemented to allow users to have the autonomy to decide the range of MC days. <br><br> |
+ | |||
+ | ==Derived variable - Hours worked== | ||
+ | <br>[[File:hm2_24.png | 350px]] <br><br> | ||
+ | To find the number of hours a victim has worked, the number of hours worked is calculated by taking the time difference between 'Supposed start work timing' and 'Accident Time'. The new variable was created using the 'Date Difference' function in JMP as shown above. | ||
+ | |||
+ | ==Binning variable - Victim's Age== | ||
+ | As the value of 'Victims' Age, spans from 16 to 82, the variable was further binned into the following categories: <br> | ||
+ | <b>Young</b>: Age 34 and below <br> | ||
+ | <b>Middle Age</b>: Age 35 to 64<br> | ||
+ | <b>Elderly</b>: Age 65 and above<br><br> | ||
+ | |||
+ | =Analysis= | ||
+ | ==Q1: Is there a correlation between the experience of employees and workplace injury? Does age plays a part too?== | ||
+ | The variables used to answer Q1 are:<br> | ||
+ | 1. Months of employment<br> | ||
+ | 2. Victim's age<br> | ||
+ | 3. Age Group (Derived variable)<br><br> | ||
− | + | [[File:hm2_4.PNG | 800px]] <br><br> | |
− | + | Looking at the above distribution of victims by their months of employment, it is apparent that the data is right skewed, where 40% of victims’ employment is less than 12 months. This means that experience in the workplace environment do play in a part in their injuries, where people who are less experienced are more prone to workplace injury as compared to the more experienced employees. <br><br> | |
− | [[File:hm2_4.PNG | | + | [[File:hm2_6.PNG | 800px]] <br> |
− | Looking at the above distribution of victims by their months of employment, it is apparent that the data is right skewed, where 40% of victims’ employment is less than 12 months. This means that experience in the workplace environment do play in a part in their injuries, where people who are less experienced are more prone to workplace injury as compared to the more experienced employees. <br> | + | [[File:hm2_5.PNG | 300px]] <br><br> |
− | [[File: | + | The distribution of victims by age reveals that most of the victims are actually in their mid 20s to mid 30s. The age was further binned to look at the proportion of victims by age group. Elderly (age 65 and above) made up the smallest proportion of victims, which indicates that being old does not mean that they are “clumsier” or less vigilant in the workplace. The younger employees (age 34 and below) actually made up 47% of the proportion of victims, and this may be due to the lack of experience in the workforce. <br><br> |
− | [[File: | + | [[File:hm2_7.PNG | 600px]] <br><br> |
− | The distribution of victims by age reveals that most of the victims are actually in their mid 20s to mid 30s. The age was further binned to look at the proportion of victims by age group. Elderly (age 65 and above) made up the smallest proportion of victims, which indicates that being old does not mean that they are “clumsier” or less vigilant in the workplace. | ||
− | The younger employees (age 34 and below) actually made up 47% of the proportion of victims, and this may be due to the lack of experience in the workforce. <br> | ||
− | [[File:hm2_7.PNG | 600px]] <br> | ||
Looking at the months of employment of the younger employees, half of them are actually new to the workplace, with employment less than 1 year. <br><br> | Looking at the months of employment of the younger employees, half of them are actually new to the workplace, with employment less than 1 year. <br><br> | ||
==Q2: What are the most common type of accident and injury in each industry, and how serious are the injuries?== | ==Q2: What are the most common type of accident and injury in each industry, and how serious are the injuries?== | ||
− | [[File:hm2_8.PNG | | + | The variables used to answer Q2 are:<br> |
+ | 1. Major Industry<br> | ||
+ | 2. Sub Industry<br> | ||
+ | 3. Accident Type Desc<br> | ||
+ | 4. No. of MC days<br> | ||
+ | 5. LOG <No. of MC days> (Transformed variable)<br> | ||
+ | 6. Nature of Injury<br> | ||
+ | 7. Cause<br> | ||
+ | 8. Injured when performing official work duties <br><br> | ||
+ | |||
+ | [[File:hm2_8.PNG | 700px]] <br><br> | ||
The above visualisation looks at the breakdown of victims by the Major industry (Manufacturing, construction, others), sub industry and the type of accident. The size of the treemap reflects the number of victims, whereas the colour reflects the seriousness of the accident which was determined by the median of the transformed variable, number of MC days (Log(No of MC days)), to account for the skewness of the variable. The purpose of the visualisation is to know the common accidents in the industries and to understand how serious those injuries are, because if there is a high number of victims who actually sustained serious/ major injuries, then there may be a need to look into the workplace safety of that industry or company. For example if majority of the victims’ injuries are caused by the use an equipment, actions could be taken to assess the risk of the equipment or probably ban the use of it.<br><br> | The above visualisation looks at the breakdown of victims by the Major industry (Manufacturing, construction, others), sub industry and the type of accident. The size of the treemap reflects the number of victims, whereas the colour reflects the seriousness of the accident which was determined by the median of the transformed variable, number of MC days (Log(No of MC days)), to account for the skewness of the variable. The purpose of the visualisation is to know the common accidents in the industries and to understand how serious those injuries are, because if there is a high number of victims who actually sustained serious/ major injuries, then there may be a need to look into the workplace safety of that industry or company. For example if majority of the victims’ injuries are caused by the use an equipment, actions could be taken to assess the risk of the equipment or probably ban the use of it.<br><br> | ||
− | Through the visualisation, we can understand that the most common accident in the construction industry is being struck by moving object. Although that is the most common accident, the injury sustained is not very severe as seen from the colour intensity. | + | Through the visualisation, we can understand that the most common accident in the construction industry is being struck by moving object. Although that is the most common accident, the injury sustained is not very severe as seen from the colour intensity.Although there are a few areas with dark shades of red which represents serious injury, there is actually just a small number of victims who are involved. <br><br> |
− | Although there are a few areas with dark shades of red which represents serious injury, there is actually just a small number of victims who are involved. <br> | + | [[File:hm2_9.PNG | 900px]] <br><br> |
− | [[File:hm2_9.PNG | | + | An example could be seen from the above visualisation where there is only 1 victim who is injured from the collapse of cranes in the construction industry, although the severity of the injury is high. <br><br> |
− | An example could be seen from the above visualisation where there is only 1 victim who is injured from the collapse of cranes in the construction industry, although the severity of the injury is high. <br> | + | [[File:hm2_10.PNG | 1000px]] <br><br> |
− | [[File:hm2_10.PNG | | + | We can also further understand the injuries sustained from the accident, and also the cause of accident, whether its self-induced or external, and if victims are actually injured when performing their official work duties. The colour intensity further shows the seriousness of each injury. <br><br> |
− | We can also further understand the injuries sustained from the accident, and also the cause of accident, whether its self-induced or | + | |
==Q3: Which are the top 3 companies that has the highest number of victims?== | ==Q3: Which are the top 3 companies that has the highest number of victims?== | ||
− | [[File:hm2_19. | + | The variables used to answer Q3 are:<br> |
− | The above table only focus on the top 3 companies with the highest number of victims. Majority of the companies has only 1 or 2 victims, however the above companies are exceptions where there is high number of accidents. Thus, we shall look into these companies to understand the victims and the injuries sustained. <br> | + | 1. Informant's company name<br> |
− | [[File: | + | 2. Months of employment <br> |
− | For example, by selecting the company, EM0187, which has the highest number of victims, we can understand that those victims are generally younger. | + | 3. Victim's Age<br> |
+ | 4. Age Group (Derived variable)<br> | ||
+ | 5. No. of MC days<br> | ||
+ | 6. LOG <No. of MC days> (Transformed variable)<br> | ||
+ | 7. Nature of Injury<br> | ||
+ | 8. Cause<br> | ||
+ | 9. Injured when performing official work duties <br><br> | ||
+ | |||
+ | [[File:hm2_19.png | 300px]] <br><br> | ||
+ | The above table only focus on the top 3 companies with the highest number of victims. Majority of the companies has only 1 or 2 victims, however the above companies are exceptions where there is high number of accidents. Thus, we shall look into these companies to understand the victims and the injuries sustained. <br><br> | ||
+ | [[File:hm2_23.png | 1100px]] <br><br> | ||
+ | For example, by selecting the company, EM0187, which has the highest number of victims, we can understand that those victims are generally younger, thus more supervision and guidance may be given to these youngsters. In this company, although it has the highest number of victims, majority injuries were quite mild, where the most serious injury was cut bruises, and was given 1 month MC. <br><br> | ||
=Dashboard visualisation= | =Dashboard visualisation= | ||
− | + | Below is the dashboard visualisation which is created from Tableau can also be accessed live from [https://public.tableau.com/profile/teo.hui.min#!/vizhome/IS428MA2/MA2/ here].<br> | |
− | + | <br>[[File:hm2_final.PNG | 1200px]] <br><br> | |
+ | |||
+ | ==Dashboard interactions== | ||
+ | <br>[[File:hm2_12.PNG | 500px]] <br><br> | ||
+ | The dashboard also allows for interaction between the individual visualisations and the possible interactions between the visualisations is illustrated above. For example, by clicking on the distribution of victims with less than 12 months of experience will reflect the breakdown of age group of these victims, severity of injuries they sustained and the cause and type of injuries of those victims. <br><br> | ||
==Summary statistics== | ==Summary statistics== | ||
− | [[File: | + | [[File:hm2_21.png | 600px]] <br> |
− | A summary of the number of victims by gender and major industry is presented to the users to give them a brief understanding of the workplace injury. <br><br> | + | A summary of the number of victims by gender and major industry and the top 3 companies with the highest number of victims is presented to the users to give them a brief understanding of the workplace injury. <br><br> |
+ | |||
==Global filters== | ==Global filters== | ||
− | [[File:hm2_14.PNG | | + | [[File:hm2_14.PNG | 150px]] <br><br> |
Global filters are implemented in the dashboard to allow users to focus on specific areas, such as a particular gender, major or minor injury and injuries that are sustained when working overtime. <br><br> | Global filters are implemented in the dashboard to allow users to focus on specific areas, such as a particular gender, major or minor injury and injuries that are sustained when working overtime. <br><br> | ||
− | + | ||
− | |||
− | |||
==Reset Filter== | ==Reset Filter== | ||
− | [[File:hm2_15.PNG | | + | [[File:hm2_15.PNG | 150px]] <br><br> |
A click to reset all filters was implemented to revert the visualisation back to the default. <br><br> | A click to reset all filters was implemented to revert the visualisation back to the default. <br><br> | ||
+ | |||
==Slider== | ==Slider== | ||
− | [[File:hm2_16.PNG | | + | [[File:hm2_16.PNG | 200px]] <br> |
As mentioned in the section of ‘Data Exploration’, as the number of MC days given varies, the slider allow users to only look at the victims within the range they have specified. However the changes from the slider will only be reflected in the ‘Seriousness of injury’ and ‘Nature and Cause of Injury’ visualisation. <br><br> | As mentioned in the section of ‘Data Exploration’, as the number of MC days given varies, the slider allow users to only look at the victims within the range they have specified. However the changes from the slider will only be reflected in the ‘Seriousness of injury’ and ‘Nature and Cause of Injury’ visualisation. <br><br> | ||
+ | |||
=Areas for improvement = | =Areas for improvement = | ||
− | [[File:hm2_17.PNG | | + | <br>[[File:hm2_17.PNG | 300px]] <br><br> |
− | <b>Initial question: Is there a correlation between the number of hours worked and the number of accidents. < | + | <b>Initial question:</b> Is there a correlation between the number of hours worked and the number of accidents.<br><br>However after finding out the hour difference between the accident time and supposed start work timing, some of the records resulted in negative values, and this was because the accident time was earlier than the supposed start work time. As only the accident date was provided in the data, there are actually 2 ways to interpret if the accident time was earlier than supposed start work time. <br><br> |
− | [[File:hm2_18.PNG | | + | [[File:hm2_18.PNG | 350px]] <br><br> |
Using the above record as an example, it can be interpreted as: <br> | Using the above record as an example, it can be interpreted as: <br> | ||
1. The victim started work earlier at 10 am instead of 8 pm, or<br> | 1. The victim started work earlier at 10 am instead of 8 pm, or<br> | ||
− | 2. The victim has worked from 8pm to the next day 10 am. <br> | + | 2. The victim has worked from 8pm to the next day 10 am. <br><br> |
− | Hence, if the date of work was provided, we could have found out the right interpretation and the question | + | Hence, if the date of work was provided, we could have found out the right interpretation and the question can then be answered. <br> |
Latest revision as of 21:53, 5 October 2016
Contents
- 1 Theme & Questions
- 2 Data Exploration
- 3 Analysis
- 3.1 Q1: Is there a correlation between the experience of employees and workplace injury? Does age plays a part too?
- 3.2 Q2: What are the most common type of accident and injury in each industry, and how serious are the injuries?
- 3.3 Q3: Which are the top 3 companies that has the highest number of victims?
- 4 Dashboard visualisation
- 5 Areas for improvement
Theme & Questions
The dataset used for this assignment is the Workplace Injury in 2014. The focus of this assignment is to look for correlation between the injuries and the other variables, and looking into the injuries of 3 industries (Manufacturing, Construction & Others) from various viewpoint. Through the data discovery, it aims to answer the following questions:
1. Is there a correlation between the experience of employees and workplace injury? Does age plays a part too?
2. What are the most common type of accident and injuries in each industry and how serious are the injuries?
3. Which are the top 3 companies with the highest number of victims?
Data Exploration
Transformed variable - No. of MC days
During the data exploration, it was discovered that some of the variables are skewed. One of the variables is ‘No. of MC days’, hence it was transformed by taking ‘LOG’.
However the range for this variable is 1 to 3600 days which actually does not make sense. According to the MOM website, the maximum paid hospitalisation leave is only 60 days. Thus there may be a case of human error in the data entry. However the number of MC that employees are entitled varies, thus these anomalies are not being removed from the data. A filter will be implemented to allow users to have the autonomy to decide the range of MC days.
Derived variable - Hours worked
To find the number of hours a victim has worked, the number of hours worked is calculated by taking the time difference between 'Supposed start work timing' and 'Accident Time'. The new variable was created using the 'Date Difference' function in JMP as shown above.
Binning variable - Victim's Age
As the value of 'Victims' Age, spans from 16 to 82, the variable was further binned into the following categories:
Young: Age 34 and below
Middle Age: Age 35 to 64
Elderly: Age 65 and above
Analysis
Q1: Is there a correlation between the experience of employees and workplace injury? Does age plays a part too?
The variables used to answer Q1 are:
1. Months of employment
2. Victim's age
3. Age Group (Derived variable)
Looking at the above distribution of victims by their months of employment, it is apparent that the data is right skewed, where 40% of victims’ employment is less than 12 months. This means that experience in the workplace environment do play in a part in their injuries, where people who are less experienced are more prone to workplace injury as compared to the more experienced employees.
The distribution of victims by age reveals that most of the victims are actually in their mid 20s to mid 30s. The age was further binned to look at the proportion of victims by age group. Elderly (age 65 and above) made up the smallest proportion of victims, which indicates that being old does not mean that they are “clumsier” or less vigilant in the workplace. The younger employees (age 34 and below) actually made up 47% of the proportion of victims, and this may be due to the lack of experience in the workforce.
Looking at the months of employment of the younger employees, half of them are actually new to the workplace, with employment less than 1 year.
Q2: What are the most common type of accident and injury in each industry, and how serious are the injuries?
The variables used to answer Q2 are:
1. Major Industry
2. Sub Industry
3. Accident Type Desc
4. No. of MC days
5. LOG <No. of MC days> (Transformed variable)
6. Nature of Injury
7. Cause
8. Injured when performing official work duties
The above visualisation looks at the breakdown of victims by the Major industry (Manufacturing, construction, others), sub industry and the type of accident. The size of the treemap reflects the number of victims, whereas the colour reflects the seriousness of the accident which was determined by the median of the transformed variable, number of MC days (Log(No of MC days)), to account for the skewness of the variable. The purpose of the visualisation is to know the common accidents in the industries and to understand how serious those injuries are, because if there is a high number of victims who actually sustained serious/ major injuries, then there may be a need to look into the workplace safety of that industry or company. For example if majority of the victims’ injuries are caused by the use an equipment, actions could be taken to assess the risk of the equipment or probably ban the use of it.
Through the visualisation, we can understand that the most common accident in the construction industry is being struck by moving object. Although that is the most common accident, the injury sustained is not very severe as seen from the colour intensity.Although there are a few areas with dark shades of red which represents serious injury, there is actually just a small number of victims who are involved.
An example could be seen from the above visualisation where there is only 1 victim who is injured from the collapse of cranes in the construction industry, although the severity of the injury is high.
We can also further understand the injuries sustained from the accident, and also the cause of accident, whether its self-induced or external, and if victims are actually injured when performing their official work duties. The colour intensity further shows the seriousness of each injury.
Q3: Which are the top 3 companies that has the highest number of victims?
The variables used to answer Q3 are:
1. Informant's company name
2. Months of employment
3. Victim's Age
4. Age Group (Derived variable)
5. No. of MC days
6. LOG <No. of MC days> (Transformed variable)
7. Nature of Injury
8. Cause
9. Injured when performing official work duties
The above table only focus on the top 3 companies with the highest number of victims. Majority of the companies has only 1 or 2 victims, however the above companies are exceptions where there is high number of accidents. Thus, we shall look into these companies to understand the victims and the injuries sustained.
For example, by selecting the company, EM0187, which has the highest number of victims, we can understand that those victims are generally younger, thus more supervision and guidance may be given to these youngsters. In this company, although it has the highest number of victims, majority injuries were quite mild, where the most serious injury was cut bruises, and was given 1 month MC.
Dashboard visualisation
Below is the dashboard visualisation which is created from Tableau can also be accessed live from here.
Dashboard interactions
The dashboard also allows for interaction between the individual visualisations and the possible interactions between the visualisations is illustrated above. For example, by clicking on the distribution of victims with less than 12 months of experience will reflect the breakdown of age group of these victims, severity of injuries they sustained and the cause and type of injuries of those victims.
Summary statistics
A summary of the number of victims by gender and major industry and the top 3 companies with the highest number of victims is presented to the users to give them a brief understanding of the workplace injury.
Global filters
Global filters are implemented in the dashboard to allow users to focus on specific areas, such as a particular gender, major or minor injury and injuries that are sustained when working overtime.
Reset Filter
A click to reset all filters was implemented to revert the visualisation back to the default.
Slider
As mentioned in the section of ‘Data Exploration’, as the number of MC days given varies, the slider allow users to only look at the victims within the range they have specified. However the changes from the slider will only be reflected in the ‘Seriousness of injury’ and ‘Nature and Cause of Injury’ visualisation.
Areas for improvement
Initial question: Is there a correlation between the number of hours worked and the number of accidents.
However after finding out the hour difference between the accident time and supposed start work timing, some of the records resulted in negative values, and this was because the accident time was earlier than the supposed start work time. As only the accident date was provided in the data, there are actually 2 ways to interpret if the accident time was earlier than supposed start work time.
Using the above record as an example, it can be interpreted as:
1. The victim started work earlier at 10 am instead of 8 pm, or
2. The victim has worked from 8pm to the next day 10 am.
Hence, if the date of work was provided, we could have found out the right interpretation and the question can then be answered.