Feesable

• GIF

  highlighting search capabilities, finding transfer fees for fords to montana

• GIF

  demo showing how visualisation (transfer fee bar and map arrow colors) change as a function of estimated transfer fee

• 

  static screenshot of app

• GIF

  highlighting the fees for a variety of cars to montana

Inspiration

When you're buying a car online, you know that it can be shipped in, but from where? And what gravity does that have on the purchase of your car? In a word: is it feesable? We wanted to answer this question.

What it does

Shows you what the transfer fee looks like for car purchases

How I built it

we built it with python and git. Rachel scraped some data (manually) from the car max website first with beautiful soup. Then, Brian read in csv data into the driver.py file. Andrew started working on linear regressions to approximate transfer fee. Rachel made a rough mock-up of the GUI in Tkinter. Andrew and Brian tried to scrape data from CarMax's private API with urllib. Afterwards, they got the plot working for the linear regression and added arrow drawings to the map, got the list of cars to display, and added the values for state codes in the menubar.

Challenges I ran into

Andrew: could not access CarMax through python urllib, ended up just using a manually collected sample. Brian: regex Rachel: Tkinter. scraping from CarMax is also hard because there is only a private API, and it was easiest to scrape an offline sample. However, it occurred to me retroactively that the smartest thing to do was probably pull data from the API and then get shipping data.

Accomplishments that I'm proud of

andrew: eating numbers (linear regression) brian: rachel: merge conflict resolution. also the guis pretty rockin'. remembered Pythagorean theorem

What I learned

andrew: brian: rachel: UI is just finding templates and themes and frameworks and then using them. so easy.

What's next for Transfer Fee Analysis for Cars

We would like to harvest a better dataset, and/or to pull data straight from the site/scrap from URLs of each car. In conjunction, we would also like the data set to include all CarMax locations, and not just generalize "location" by state. We also want to implement a histogram display by location and shipping cost to help the user better understand the data and options for car purchases. Additionally, we also want to filter for cars by criteria, display all viable options on the map, and more ways to compare car options.

Built With

• beautiful-soup
• math
• matplotlib
• python
• tkinter
• tkk