IBM Capstone Project | Car Accident Severity Seattle, USA | Coursera Community

IBM Capstone Project | Car Accident Severity Seattle, USA

  • 20 October 2020
  • 6 replies
  • 775 views

Userlevel 2
Badge +1

Hi community,

I would like to share with you my insights that I gained in course of my capstone project, working with car accident data from Seattle, USA.

 

The purpose of this exploratory data analysis exercise is to assess the possibility and accuracy to predict car accident severity in Seattle, USA by means of supervised machine learning models, exploiting collision track records from past accidents that were recorded by the Seattle Police Department (SPD) and provided as open data by Traffic Records.

Being able to predict car accident severity from extrenal factors like weather, location, road conditions as well as speeding, influence of alcohol/drugs etc. will allow the government to put appropriate meassures in place to reduce accident severity, but above all, allow the police and first response teams to channel their resources and increase efficiency.

Using car accident track records from March 2013, three different machine-learning methods, namely K-Nearest Neighbours (KNN), Decision Trees, and Logistic Regressors, were benchmarked against each other

While the exploratory data analysis suggests, that almost 90% of all accidents, involving pedestrians and 88% of collisions involving cyclists lead to injuries (compare: 28% of accidents without pedestrians/cyclists lead to injuries), the tested machine learning models generally had problems to correctly predict `SEVERITYCLASS=2` and therefore exhibited a high number of *false negatives* for this class which can in real life lead to a wrong allocation of resources of police and first responders and potentially end deadly.

 

The link to my Jupyter Notebook on Github can be found here:

https://github.com/martinvau/Coursera_Capstone/blob/master/src/IBM_Capstone_-_Car_Accident_Severity_Submission.ipynb

 

Hope you find it interesting and helpful for your submission!

 

Cheers,

M

 


6 replies

Badge +1

Wow ! This is an eye opener and very interesting. I’m definitely going to do more studies. Thank you Vau.

Badge

Awesome..This is how a capstone to be done..

Badge

Awesome 👍

Userlevel 1
Badge +1

Interesting Project, in what certification/course did assign you this? 

Thanks for sharing!

Userlevel 6
Badge +11

Hi community,

I would like to share with you my insights that I gained in course of my capstone project, working with car accident data from Seattle, USA.

 

The purpose of this exploratory data analysis exercise is to assess the possibility and accuracy to predict car accident severity in Seattle, USA by means of supervised machine learning models, exploiting collision track records from past accidents that were recorded by the Seattle Police Department (SPD) and provided as open data by Traffic Records.

Being able to predict car accident severity from extrenal factors like weather, location, road conditions as well as speeding, influence of alcohol/drugs etc. will allow the government to put appropriate meassures in place to reduce accident severity, but above all, allow the police and first response teams to channel their resources and increase efficiency.

Using car accident track records from March 2013, three different machine-learning methods, namely K-Nearest Neighbours (KNN), Decision Trees, and Logistic Regressors, were benchmarked against each other

While the exploratory data analysis suggests, that almost 90% of all accidents, involving pedestrians and 88% of collisions involving cyclists lead to injuries (compare: 28% of accidents without pedestrians/cyclists lead to injuries), the tested machine learning models generally had problems to correctly predict `SEVERITYCLASS=2` and therefore exhibited a high number of *false negatives* for this class which can in real life lead to a wrong allocation of resources of police and first responders and potentially end deadly.

 

The link to my Jupyter Notebook on Github can be found here:

https://github.com/martinvau/Coursera_Capstone/blob/master/src/IBM_Capstone_-_Car_Accident_Severity_Submission.ipynb

 

Hope you find it interesting and helpful for your submission!

 

Cheers,

M

 

Hello @MartinVau 

Thanks for sharing your invaluable experience and knowledge acquired by working on the Capstone Project work.

Regards,

Saheli

Badge +1

Hi Martin,

thanks for sharing. I’ll have a look at your work; meanwhile, if you are interested, you can have a look at my project on the same subject (Car Accident Severity Seattle, USA): https://www.linkedin.com/pulse/alert-today-alive-tomorrow-predicting-road-safety-sebastiano-denegri/?trackingId=OGz4vs40QDuo2FjnNHO0Xg%3D%3D (link to the notebook: https://nbviewer.jupyter.org/github/SebastianoDenegri/Coursera_Capstone/blob/master/applied_data_science_capstone-final_version_geocoord.ipynb).

I’d actually created 2 models:

  • Model 1 - Classification of Risk. Is the risk of getting into car accident higher than usual? The model doesn’t predict the likelihood of getting into a car accident, but it returns a binary classification able to predict the class of risk based on historical conditions, considered either not-dangerous, or dangerous (that is conditions under which the majority of accidents happen), from an objective point of view.
  • Model 2 - Classification of Collision Severity. How severe the potential collision is.

Regarding your examples (collisions involving pedestrians/cyclists... lead to injuries), I didn’t use these features to build my models, because I considered them as not available at the time of making a predictions: you don’t really know whether an accident involves pedestrians, cyclists, extra-vehicles… if the accident hasn’t taken place yet. Anyway, you may have a different perspective about. 

Looking forward to hearing your thoughts.

Cheers,

Sebastiano

 

Reply