Petrolytics

HomePage for Site
ChartJS Visualizations
HeatMap Based on Scrapped Data-Sets
Advertisement for Petrolytics
Petrolytics Logo

Petrolytics is for the Conoco Phillips Challenge

Inspiration

Data-Visualization and data-analysis is a very exciting and interesting field to work with. And we found the Conoco Phillips to be the perfect challenge for learning these types of skills.

What it does

We scrapped various data-sets to obtain information about the 3 biggest refineries and their specific city-locations in the US. Based on that data we over-layed a heat-map on the map of the United-States to gain insight on the major locations that refineries are found in.

How we built it

We scrapped the data with BeautifulSoup4 a python library that is very well suited for web-scrapping. As for the heat-map generations, we used folium which is a python library that based on data-points, will generate heatmaps. However, since we scrapped all the data the MapBox-API was used to geolocate the city names to actual longitude and latitude coordinates. For our cursory data-analytics we used chart.js to showcase some interesting data visually inside the browser. Since our project didn't make use of a backend we decided to use Google Cloud Buckets to host our site.

Challenges we ran into

Given our skillset we planned to remain focused on adhering to strictly vanilla JS and python. However, heatmaps and data-visualization libraries tend to be very complex and often require a framework. However, we were able to find libraries that suit our needs perfectly for our project, which were folium & chart.js

Accomplishments that we're proud of

This is our first project working with data-visualization and the fact that we were able to upload our project for others to interact with, makes the project pop that much more. It was also a great exercise in practicing web-scrapping and pushing the features of the python language.

What we learned

Data-Visualization with Python and JS Web-Scrapping and Cleaning Data Hosting content on Google-Cloud Interacting with API-Data in Python

What's next for Petrolytics

We were rather unsatisfied with the quality of APIs regarding Oil-refineries available. Mostly due to complexity and somewhat lacking features. We're considering building our own API based on publicly available data-sets in order to provide a more seamless workflow.

In terms of features, we'd love to perform predictive insights with Machine-Learning based on our data. And to also dynamically prepare visualization for greater interactivity and use.