Brainstorming!!

[rasmusson]

I have a lot of data and I don’t know what to do with it.

In Oslo where I live we have a bike sharing system called Oslo bysykkel. The system is built up of a number of stations with bikes. If you have an subscription you can pick up a bike and you have to return it to any of the stations within two hours.

Oslo bysykkel has an API that gives the number of available bikes and available locks at a station. For the last three month I have collected the this info for all stations every five minutes and stored it in a database.

Does anyone have an idea of what to do with this data? Cool data visualizations? Corelation with other data, weather maybe? Predictions? Any and all ideas are appreciated.

I also have an android app that shows the status for the stations on a map. I would love to use this data in some way in the app.

 

[bugman]

First, do you have any specific questions relating to these data?

Next, which software packages are you familiar with and have access to?

There are a number of cool spatial things that you could do with this in R, GIS, Tableau for example, but before I provide links or info. Let me know waht you are savvy with.

 

[rasmusson]

First, do you have any specific questions relating to these data?

Mostly I just wanted to have ideas on what one could do with the data, interesting conclusions one could make. I have one specific question relating to my app. "what is the probability of there being an available bike at a specific station in the next 10 minutes?" or "how long before there will probably be a bike available at a specific stand?". I would think whether will have a big impact so this have to be taken into account.

Next, which software packages are you familiar with and have access to?

None what so ever I have no experience with statistics.

There are a number of cool spatial things that you could do with this in R, GIS, Tableau for example, but before I provide links or info. Let me know waht you are savvy with.

Ok sounds awesome thanks

 

[rasmusson]

Did you see my response?

 

[noetsi]

If you have no experience with statistics anything complicated is not a good idea. Because 1) it won't be simple to do and 2) you might not understand what you get or the limitations on it. Descriptive statistics is probably a good place to start and can show interesting results. For example you can show bike pick ups by day and month (showing which days are more heavily used and possibly patterns within the day). You might then seek to discover what this entails for the city (if bike usage is heavier at certain times it might influence traffic patterns for example).

Generally I would suggest time series analysis or queing models (which evaluates demand relative to supply of services in part) but if you have a limited background in these and no computer software to use them that would be a tall order to do.

 

[hlsmith]

I have a completely different idea. I would contact the Public Health College at a local university and get in contact with graduate students required to have a masters project or dissertation.

I would work with them and come up with questions you want answered along with their own questions and have them churn that data up every which way. Make them do the work and literature reviews.

You could look to other programs as well if interested, such as civil engineers, economic students, etc. This way you can have them engage in countless hours analyzing your data and they can graduate and perhaps have a project that they enjoy. I know an environment health student that looked at traffic patterns and acccidents in bicyclists for their dissertation.