jansenkat / scooter-app

Scooters in Austin

Data Analysis and Visualization Boot Camp 2019-2020 JSON& Team (2) (T/Th)

If you've been out and about on campus or in downtown Austin, you can't help but notice that scooters are all over the place. We decided to examine distance and duration of scooter rides and where people complain about scooters.

In our first project we merged data sets to create visualizations comparing complaints to rides, to see if there were some neighborhoods that were being the squeaky wheels or if there really was a scooter problem in certain areas. As we had hypothesized, scooters that find their way into further out neighborhoods are less likely to be picked up and thus take longer to be ridden back into town, are much more likely to result in a complaint.

In our previous presentation, we called out some limitations and remaining questions that we had, and have demonstrated some of those here in Project 2.

In this project we used mySQL, FastAPI, Uvicorn, Plotly and Cloud Run for this web app.

Our Project is a dashboard for reviewing complaints, rides, and certain trends that we found interesting. The routes we decided to demonstrate for you all are:

• Longest rides

• Rides that went nowhere

• Explore a random ride that someone took in Austin

• What is the red zone, or neighborhoods you just shouldn't leave scooters in?

• What are the percent of complaints, rides in, rides out, and other stats about specific zip codes?

We hope you'll find this data to be as interesting as we did, regardless of whether you consider yourself a part of the shared mobility revolution or not.

scooter-app's Issues

random api route

/random_api Route returning 500 Internal Server Error

CSS Home Page

Create CSS for home page

Create Jinja template

Queries

create sqlalchemy queries to pull data into plotlyjs files

create query specific views

Postman Testing

Must use Postman with at least one request for each route
export a collection and share in github

RedZone Plot

Create RedZone plot

Longest Ride

Start/End Address on Map

Presentation

organize / script presentation
anticipate questions

Add error handling

At least to plot files in case api routes don't return data

Deployment

Deploy with heroku
link heroku to db?

Zero Distance/Nowhere

Zero Distance Initial Load does not display Number of Trips
----- Other Dropdown options show high load time / freeze page

10 most common routes

Home Route Test

Homepage HTML template Test
Test the app with Thomas' template

• Run the fast api app with the template
• Seeing if API works with the template

README.md

Create and update Readme to be used in presentation as well as guide for repo.

Scooter Trips Route no longer exists

We had this initially as an api endpoint. Did we wanna replace w/ sometjing else?

API to Database Connection Test

Test that the API is connected to database
• 2 API routes (one for each table)

• Test API routes (testing functionality of fast API)
• Docs page should be automatically created
• Should return a JSON of the table (one table or the other)

Database

Pull database form google cloud to thumbdrives

Draft Proposal

Create basic proposal doc

complaint map

0 Distance Rides

Random Scooter Plot

Zipcode analytics

HTML Home Page

Create html file for home page
Include endpoints and how to use them (on main page or linked page)

Front end differences on deployed app

• Deployed App does not "look" the same as when running locally

Create App file

At least 5 routes
Linked with database (Jason's google cloud db)

Longest Ride Page

Longest Ride - Duration Dropdown.click displays only 5 traces
----- Toggling back to Distance on Dropdown does not work/refresh plot

Zipcode Test

Run the URL/zipcode
Test If zip code alone will work with Jinja

• Page should display numeric zip code that was entered in URL

Pipenv

Create Pipfile
should include (versions to be specified):

• python
• sqlalchemy
• flask
• flask-sqlalchemy
• plotly
• pandas
• black --pre --dev

query optimization