Add fare metrics from NTD using the new transit class

.gitignore

@@ -6,3 +6,10 @@ package-lock.json
 setup/
 miniconda.sh
 test_js*
+shapefile
+*.png
+*.pdf
+data/
+*chloropleth_map.html
+*.gif
+denver/

scripts/otp/Makefile

@@ -0,0 +1,41 @@
+town := denver
+
+
+.PHONY: otp
+
+
+otp:
+	# create directory for data and config
+	-mkdir $(town)
+	cp build-config.jsonc $(town)/build-config.json
+	# download OSM
+	if [ ! -f $(town)/osm.pbf ]; then \
+		curl -C - -L -k https://download.geofabrik.de/north-america/us/colorado-latest.osm.pbf -o $(town)/osm.pbf; \
+	fi
+	#
+	# past logic for manual GTFS downloading.
+	# this is superceded by the scrape ipynb
+	#
+	# download GTFS
+	# this is the latest gtfs file
+	#curl -L -k https://www.rtd-denver.com/files/gtfs/google_transit.zip -o $(town)/denver-rtd-gtfs.zip
+	# 2022
+	#curl -C - -L -k -H "Accept: application/zip" https://web.archive.org/web/20220308061530/https://www.rtd-denver.com/files/gtfs/google_transit.zip -o $(town)/denver-2022-rtd-gtfs.zip
+	#curl -L -k https://web.archive.org/web/20220308061530if_/https://www.rtd-denver.com/files/gtfs/google_transit.zip -o $(town)/denver-2022-rtd-gtfs.zip
+	#
+	#
+	#
+	# build graph and save it onto the host system via the volume
+	@echo "Go to Docker Desktop and change resource to maximum memory and swap."
+	docker run --rm \
+		-e JAVA_TOOL_OPTIONS='-Xmx32g' \
+		-v ./$(town):/var/opentripplanner docker.io/opentripplanner/opentripplanner:latest --build --save
+	# load and serve graph
+	docker run -it --rm -p 9999:8080 \
+		-e JAVA_TOOL_OPTIONS='-Xmx32g' \
+		-v ./$(town):/var/opentripplanner \
+		docker.io/opentripplanner/opentripplanner:latest --load --serve
+
+manual:
+	java -Xmx2G -jar otp-2.6.0-shaded.jar --build --serve /home/username/otp
+	# osmium extract --bbox=-123.043,45.246,-122.276,45.652 -o portland.pbf oregon-latest.osm.pbf

scripts/otp/README.md

@@ -0,0 +1,116 @@
+# Time and Cost Solution
+
+launch docker container according to the readme in
+https://github.com/e-mission/em-public-dashboard
+
+Here is an overview of the objective.
+
+![Overview diagram](costtime.svg)
+
+## How can I achieve choice model analysis with historical data?
+
+The most crucial step in the process is finding and retrieving the
+data. Quality data that is complete (spanning many years)
+and accurate will allow for a more informed model.
+
+## Time
+
+We want to know the time it takes for an individual to use a particular mode
+to get to a destination
+
+### Transit
+
+Transit is nuanced since there are many modes such as bus, rail, and tram.
+
+#### Data Sources
+
+GTFS (transit data standard) sources are varied, with some requiring monetary payment
+
+| **Website Name**                            | **Dates**           | **URL**                                                                 | **Paid**                    | **Notes**                  |
+|---------------------------------------------|---------------------|-------------------------------------------------------------------------|-----------------------------|----------------------------|
+| Transitland                                 | 2016-Present        | [https://www.transit.land/feeds/f-9xj-rtd#versions](https://www.transit.land/feeds/f-9xj-rtd#versions) | Yes Unless Edu Academic Plan          | Does not have historical data without paid plan          |
+| transitfeeds AKA OpenMobilityData           | 2015-Dec 2023       | [https://transitfeeds.com/p/rtd-denver/188?p=26](https://transitfeeds.com/p/rtd-denver/188?p=26) | No       | API Deprecated             |
+| Scrape of transitfeeds AKA OpenMobilityData | Up to March 2021    | [https://raw.githubusercontent.com/interline-io/scrape-of-transitfeeds/refs/heads/master/json-scrape/feedVersionS3Urls.csv](https://raw.githubusercontent.com/interline-io/scrape-of-transitfeeds/refs/heads/master/json-scrape/feedVersionS3Urls.csv) | No                          |  Does not have recent data                 |
+| Mobility Database                           | Feb 2024-Present    | [https://mobilitydatabase.org/feeds/mdb-178](https://mobilitydatabase.org/feeds/mdb-178) | No                          |  Does not have historical data            |
+| Wayback Machine                             | Sporadic            | [http://web.archive.org/web/20240000000000*/http://go-rts.com/gtfs/google_transit.zip](http://web.archive.org/web/20240000000000*/http://go-rts.com/gtfs/google_transit.zip)              | No | Can be used to fill in gaps between the other sources
+
+The CanBikeCO dataset does not go into 2024.
+
+Thus, the most easy option was to scrape transitfeeds. This was done previously by someone else,
+but that scrape was too old. Therefore, `scrape.ipynb` allows for an automated retrieval
+of all GTFS data, for example, RTD Denver.
+
+With this `scrape.ipynb` notebook, we are able to receive the astonishingly vast GTFS historical data from 2015 to the end of 2023. Datasets sometimes were released multiple times on a single
+day, so it seems that every single change that the agencies made were recorded by transitfeeds.
+Small changes such as the relocation of a singular stop would still reissue all of the data that remained unchanged, introducing redundancies. A form of version control, if implemented as a standard, would stop this.
+
+#### GTFS Combination
+
+It is not needed to have GTFS for every single day; it is unrealistic, as well, for computation
+analysis to churn such significant amounts of data. We used only 1 GTFS file for every quarter
+of the year. To distill this dataset, use the Jupyter cell marked `# Distill GTFS` in the
+`scrape.ipynb`.
+
+Then after distilling, download from 
+https://github.com/OneBusAway/onebusaway-gtfs-modules/blob/master/docs/onebusaway-gtfs-merge-cli.md
+and change the version accordingly within the jar filename in the following command.
+
+```bash
+java -jar -Xmx16G ~/Downloads/onebusaway-gtfs-merge-cli-3.2.4.jar denver/*.zip outputCombined.zip
+```
+
+#### OpenTripPlanner Launch
+
+Ensure that docker is installed. This is not tested on Windows, so we assume you
+have `make` available to run the Makefile. OTP does not work on Windows anyways.
+
+The Makefile downloads the geofabrik pbf file for Colorado. This can be
+parametrized, as can the `scrape.ipynb` notebook into a Python class/functions.
+
+```bash
+make
+```
+
+#### Transit Time Python API
+
+Once OTP is loaded, verify that it works within the browser by going to
+localhost:9999
+
+The routing should work with dates several years in the past e.g. 2022
+but some years in the data distilling cell skipped years such as 2015.
+
+Navigate to the Jupyter cell prefaced with `# GraphQL Interface` to have
+a Python API for getting transit times.
+
+
+## Cost
+
+### Transit
+
+It is true that GTFS provides fare information.
+However, agencies tend not to provide it. But NTD has all agencies' information.
+
+Fare data for transit is taken from NTD. The average fare for each particular agency
+is calculated for any arbitrary person- take, for instance, Gainesville, Florida.
+Most riders are students who attend the University of Florida, and all of their
+rides are free (this is the same for any faculty and staff).
+
+There are still individuals who pay for fare. So Gainesville's average is $0.07
+according to NTD.
+
+To calculate average fare, use the e-mission-common function:
+
+```python
+# gainesville fl coordinates for the bus (thats why mode is MB)
+# longitude, then latitude.
+# google maps prefers latitude, then longitude.
+# otherwise you will be in antarctica
+intensities = await emcmft.get_transit_intensities(2022, [-82.328132, 29.626142], modes=['MB'])
+
+# Access and print the average_fare
+average_fare = intensities[0]['average_fare']
+print("Average Fare:", average_fare)
+# 0.07352028470412668
+# wow, seven cents! $0.07
+```
+