Pedestrian-and-bike-crashes

A Twitter bot that tweets Bicyclist and Pedestrian related crashes, using Citizen data.

Run

npm install to fetch and install dependencies.

To run: node index.js --location cityNameOne, where cityNameOne is the name of the city you want to fetch and tweet data for (see keys section for more information).

Options

By default, this code will only tweet out events that happened within one day. You can optionally tweet data for more than 1 day by including option --days in your node command. File tweetIncidentSummaries.json will store summaries of all crashes that you grab from citizen.

Keys file

To use, create a keys.js with the following format for each city you want to fetch and tweet data for:

module.exports = {
  cityNameOne: {
    consumer_key: 'consumer_key',
    consumer_secret: 'consumer_secret',
    access_token: 'access_token',
    access_token_secret: 'access_token_secret',
    lowerLatitude: 'lowerLatitude',
    lowerLongitude: 'lowerLongitude',
    upperLatitude: 'upperLatitude',
    upperLongitude: 'upperLongitude',
    timeZone: 'America/New_York',
  },
  cityNameTwo: {
    consumer_key: 'consumer_key',
    consumer_secret: 'consumer_secret',
    access_token: 'access_token',
    access_token_secret: 'access_token_secret',
    lowerLatitude: 'lowerLatitude',
    lowerLongitude: 'lowerLongitude',
    upperLatitude: 'upperLatitude',
    upperLongitude: 'upperLongitude',
    timeZone: 'America/Los_Angeles',
  },
  ...
};

To get the consumer_key, consumer_secret, access_token, and access_token_secret values, you'll need to sign up for a Twitter Developer account. When you first sign up for an account, Twitter will give you keys. Ignore these, because they don't have the configuration that they will need. This is what you will need to do:

1. Make a new Project App
2. Give that app Read and Write permissions
3. For 'Type of app', select Web App, Automated App, or Bot
4. In 'App Info' set the Callback URI to http://localhost
5. In 'App Info' set the Website URL to your app's url, for example https://twitter.com/rvacrashbot
6. Save these settings
7. Go to the 'Keys and Tokens' tab
8. Generate new Consumer Keys: API key and Secret then place in the keys.js file
9. Generate new Authentication Tokens: Access Token and Secret then place in the keys.js file

To get the lowerLatitude, lowerLongitude, upperLatitude, and upperLongitude values, you'll need to go to the Citizen explore page. Open up the developer tools in your browser, go to the network tab, and select the city you want to see incidents for. Select the trending API call and the latitude and longitude values are the parameters in the URL.

The timeZone value is used for converting incident times (Citizen stores them in UTC) to the city's local time. The timeZone value for your city can be found here - use the value in the TZ database name column. If your city isn't listed just use the closest major city (ex. for Philadelphia use America/New_York).

---

Advanced

Satellite image of crash site

The bot can tweet zoomed in satellite images of the crash site, in addition to the standard citizen map. The Google map is a zoomed-in satellite image of the location of the crash and the Citizen map shows a simplified larger context. To enable inclusion of maps in your incident tweets, you'll first need to add a googleKey property to your city object in the keys.js file containing an API key from a Google Developer Platform account. You will need to enable the Maps Static API for your key. Then place your API key in the keys file:

module.exports = {
  cityNameOne: {
    consumer_key: 'consumer_key',
    consumer_secret: 'consumer_secret',
    ...
      googleKey
:
'google_api_key',
},
...
}
;

Then, call the bot with the --tweetSatellite flag and each incident thread will contain both a satellite image of the crash site and the standard citizen map:

node index.js --location atlanta --tweetSatellite

Representatives

The bot has the ability to determine the exact city council district a crash occurred in, and tweet the name of the district and representative in addition to the crash information.

To set this up, you'll need to:

1. Create a geojson file of your city council districts
2. Host the geofile on the internet
3. Create an object in the representatives.js file for your city (atlanta is included as an example)
4. Run the bot: node index.js --location cityNameOne --tweetReps

Create a geojson file of your city council districts

A geojson file is a file containing geographic shapes. The bot uses this file to map a crash to the city council district it happened in.

You can create a geojson file of your city council districts by going to geojson.io and tracing out the districts. You'll also need to add a NAME property to each district you draw.

Host the geofile on the internet

Once you've created the geofile, you'll need to host it somewhere so the bot can download it. Examples of places you could host it are Google Drive or in an AWS S3 bucket. You'll need the URL to the file.

Create an object in the representatives.js file

Next, you'll need to add an object to the representatives.js file that maps the name of each city council district to its representative's name (or Twitter account if you want to tag them). You'll also need to add the geojson url and representative district term. The representative district term is just what a city council district is called in your city (for example, maybe you live in Chicago which would refer to city council districts as wards).

At Large city councilmembers

If your city has at large councilmembers, you can include an atLarge property with an array containing their names (or Twitter usernames). The bot will tweet their names along with the number of crashes that occurred over the last 24 hours.