DATA AND AI TRACK SUBMISSION

My Name is Satyam Goyal ([email protected])
This project is my Submission for HACK-IT challenge by OCBC in DATA and AI Track.

Scope of Assignment:

• Use the US Traffic 2015 Dataset,publicly available on Kaggle, to visualise the traffic patterns.
• This assignment aims to clean and analyse the dataset, create appropriate models and visualise them using the proper software.

Functionality:

• Use the appropriate algorithms and models to find out the top 5 most obvious patterns from this data.
• Support your hypotheses with appropriate data.

GETTING STARTED

• Clone this repository: https://github.com/SatYu26/OCBC-Submission-Data_and_AI_Track.git
• cd into the folder: cd OCBC-Submission-Data_and_AI_Track
• Open the project in Jupyter Notebook using command jupyter notebook
• You can also open the project by opening the OCBC_Submission_Satyam_Goyal.ipynb file in Google Colab. Visit HERE and upload the ipynb file to open it in Colab.
• After opening the project change the location of Traffic Dataset in code cell/block 2 according to the location of dataset in your device.
• Run each code block and see the results.

OUTPUT

Pattern 1:

On the basis of below bar plot between Traffic count and Days of weeks across all time frames we can conclude that:

• The maximum amout of Traffic occurs on Day 6 i.e saturday around the hours of 12 to 16 i.e. Afternoon Hours.
• Also the traffic on other days is almost similar and the minimum amount of traffic occurs during eary morning hours between 1 to 4.

Pattern 2:

In the below Bar plot we have created the Graph between Direction of travel and The traffic count on Diferrent days across all time Frames.

1. With this we can Easily conclude that Northeast-Southwest Direction have the maximum amount of traffic specially on Hours 1 to 4, 4 to 8 i.e. early mornings and at Late nights between hours 20 to 24.

2. We can also conclude that The traffic on time range 8 to 12 is extremely less irrespective of Direction of travel.

We can also conclude on the basis of the below density graph that highest density of traffic is in Direction number 1 and 5 i.e. North and South across all the times included.

Using the below Graph we can further Solidify our above theory that the maximum amount of traffic occurs in the Direction of Northeast-Southwest and in time range of 13 to 16 and 17 to 20.

Pattern 3:

In the below Graph we have created the Graph between Lane of travel and The traffic count on Diferrent days across all time frames.

1. With this we can Easily conclude that Lane number 8 have the maximum amount of traffic specially on Hours 1 to 4 and 4 to 8 i.e Early Mornings, and the Least amount of traffic on Lane number 1.

2. We can also observe that on the time range of 8 to 12 the traffic is very least irrespective of Lane number.

With the below mentioned Area graph between Station ID and Lane of Travel We can conclude that:

• The Reason for Maximum Traffic on the Lane number 7 and Lane number 8 was because There are more number of stationed located on that path and because of that people Travel more across those lanes.

Pattern 4:

In the below Graph we have created the Graph between Functional Classification and The traffic count on Diferrent days across all time frames.

1. With this we can Easily conclude that Urban Principal Arterial: Interstate and Other Expressways have the maximum amount of traffic specially on Hours 1 to 4 and 4 to 8 i.e Early Mornings and Late night between hours 20 to 24, also the Least amount of traffic on Lane number 1.

2. We can also observe that on the time range of 8 to 12 the traffic is very least irrespective of Functional Classification and on the Rural Local system we have the least Traffic.

Pattern 5:

From the below mentioned Bar Graph between Functional Classification and Lane numbers we can Conclude that:

• Most number of people from Rural Areas travel mostly on Outer Lanes (i.e. 5, 6, 7, 8, 9) instead of main Lanes.

• Infact we can clearly see that There are almost negligible people from Urban Areas travelling in outer lanes except of lane 8 as there as most number of stationes situated in that lane.

We can further solidify our above mentioned argument that Most amount traffic occurs in Urban: Principal Arterial - Interstate Area in hours 13 to 16 and 17 to 20 i.e. in Evening time mostly.

Data Model:

Below i have presented the Data Model of the complete project and its final model.

Model Result:

• Our Random Forest Regressor model Performed very nicely with the accuracy of 95.24%.
• You can also Download the model from HERE.

Explain why you chose this particular model for solving the problem?

• The reason we chose this Random Forest Regressor model for solving the problem was because of the accuracy and depth this model provides to my output.

• The reason for choosing Lane of Travel as output of my model was because during Data Analysis phase, we observed a pattern in which most of the data had a similarity and good relation with Lane of Travel column and therefore it will provide us with high accuracy and consistency throughout the model.

FULL PROJECT

• To look at the full project with Dataset and Output Model Visit this Google Drive link HERE.
• To see the working Google colab file visit HERE.