Houston Science City Map

Current map is live as a github.io page

Purpose:

This is an exploration of the concept of trying to find all the different places science is created, taught, and applied in Houston and put them on a single map.

Although it is not exhaustive, it is amazing how fast you get up to several thousand points where science is done in Houston. The goal was to give people a greater appreciation of how pevarsive Science is in their their community. This was inspired by the March for Science in 2017. Although the discussion around cuts in funding and hurdles for government communication of science is centered on government agencies, those impacts will get dispursed rather quickly out to state and local agencies where science is used as well as the countless private companies who will be negatively effected if the science infrastructure in the U.S. is negatively impacted.

Don't trust the map yet, there are still false positives... maybe 20%

Possibe End Product:

Series of maps exploring the geographic distribution of sciencec in Houston

- Key science hubs only:
	- Johnson Space Space
	- Medical Center
	- Universities (Rice & Univ. Houston campuses & others?)
	- Exxon, BP, Shell, Chevron, Citgo (big names with large office complexes)
- #ActualLivingScientist hashtag tweets
	- approximate position only
	- include text & photos on hover? or click? or link to elsewhere? 
- User generate map based on science filters?
	- generate filters based on search terms used in job titles
	- generate aggregate filters based on a science ontology and job titles used in job search data collection
- All data points on one map
	- color and shapes map to?
		- schools, medical, oil&gas, aerospace, and other?
		- Science taught, created, applied 
			- (plurity wins, so Exxon is applied even though they have a lot of patents too)
	- use size to show hubs? 
- user generated context can be added to the map via integration with a google spreadsheet?
- heat map(s) to show densisty of certain types of science or industries? 
	- How accurate are our ideas or where things clustter?
- user drawn shapes
	- do they match actual distribution based on this data?
		- Can they draw where Johnson Space Center is?
		- Can they predict how distributed or clusttered a specific category is?
			- "How well do you know Science in Houston?"

non-Map data visualizations:

- job search term vs. number of resulting companies, bar chart?
- companies in order of number of different job search term matches, bar chart or tree diagram?
- science-related places on a route between two points
- street-map API photos returned to form a gallery of 'places where science is done'?

Key Data Sources:

- List of science related job titles
- Job search information from websites
- Google Places API to find lat longs associated with company names from teh job search results
- Twitter posts gathered with the twitter API to return science related hashtags within certain radius of zip code centroids (centerpoints)
- geoJSON for various categories (schools, hospitals, museums, etc.) being pulled via a get request that returns geoJSon from the city of Houston's <a href='http://cohgis-mycity.opendata.arcgis.com/' target="blank">open data GIS portal</a>.

Current Status:

At this time, I have a single map with all the job search data points as well as three geojsons for hospitals, schools, and libraries from the city of Houston's open data portal. I also have a json of various twitter posts with the #actuallylivingscitentist hashtag in the Houston area grouped around zip code centroids. I have not put the twiter data onto a map yet.

Methods (in progress):

Most of the datapoints come from the job search data. Basically I start with a list of science jobs, I look for those jobs programatically in the Houston area over the last 30 days and get back the names of all the companies advertising for that job or a very similar sounding job. I then take that list of companies and use the google places API to return all the addresses of each company in Houston. I then put those locations on the map. If clicked, a tool tip will pop up with job title used in the search. Eventually, I will use a random forest machine learning to clean up the data as there seems to be at least 20% false positives in the job search location data right now.

In Progress:

The most immediate next step is to use machine-learning to clean up all the false positives in the data. For example, satellite is often returning a satellite location of some business, not a company that works on satellites. After that, the next steps will be around making the maps more user friendly such that they can tell a story. I also need to spend some time getting the map to load faster. Right now the job search geojson loads very slowly and the user can accidentaly stop the load if they try to play with the map before it is finished loading.

justingosses / houstonsciencecity Goto Github PK

houstonsciencecity's Introduction

Houston Science City Map

Purpose:

This is an exploration of the concept of trying to find all the different places science is created, taught, and applied in Houston and put them on a single map.

Don't trust the map yet, there are still false positives... maybe 20%

Possibe End Product:

Series of maps exploring the geographic distribution of sciencec in Houston

non-Map data visualizations:

Key Data Sources:

Current Status:

Methods (in progress):

In Progress:

Current Github Location:

Started the map from this example:

Original List of Potential Map Data:

Early Map

houstonsciencecity's People

Contributors

Watchers

Recommend Projects

Recommend Topics

Recommend Org