publiclab / wherewebreathe

wherewebreathe.org
2 stars 7 forks source link

bulk anonymized download links on graphs, in footer #14

Closed jywarren closed 9 years ago

jywarren commented 10 years ago

possible anonymous aggregate database download formatting in both conventional csv/excel and custom plaintext for analysis in stata

Let's display these in a Bootstrap download popover on each graph page, and a total bulk download from footer and/or settings page

@shapironick - can you provide some example files for the stata format you're asking for? It's just text, right?

shapironick commented 10 years ago

Sounds great. Here's an example of the format Nicole would need (drop box link below). It can be an excel sheet it just needs each question to be answerable in the binary (so a multiple choice would not show just the answer selected by name but each answer with a 0 or 1 (also a period is used to designate unanswered questions as a no is different than a non-answer.

https://www.dropbox.com/s/2ocbuhrs7m4nc6g/Symptoms%20example%20.xlsx

jywarren commented 10 years ago

OK great -- presumably we can do that as CSV instead of excel as well. Then I think we may only need one format. Nick, want to try answering the Google form a few times and downloading and attaching the output here so we can see and discuss/modify how the data is presented there? Thanks. On Jun 19, 2014 9:39 AM, "shapironick" notifications@github.com wrote:

Sounds great. Here's an example of the format Nicole would need (drop box link below). It can be an excel sheet it just needs each question to be answerable in the binary (so a multiple choice would not show just the answer selected by name but each answer with a 0 or 1 (also a period is used to designate unanswered questions as a no is different than a non-answer.

https://www.dropbox.com/s/2ocbuhrs7m4nc6g/Symptoms%20example%20.xlsx

— Reply to this email directly or view it on GitHub https://github.com/publiclab/wherewebreathe/issues/14#issuecomment-46560991 .

shapironick commented 10 years ago

yes csv or ods should be a-okay.

Here is the csv of three responses: https://www.dropbox.com/s/3h7t770fc676wbh/Survey%20Draft%201%20%28Responses%29%20-%20Form%20Responses%201.csv

I think this output is great for lay users but as the cells are filled with the answers as words and not a 0, 1, or "." I'm thinking that it wouldn't work with stata. The link i shared in the last comment was data that was transcribed from a google spread sheet (from a google survey) into a stata ready format and it was pretty labor intensive even though its only 20 or so respondents.

The response summary is pretty amazing. A few tweeks will be needed but its pretty encouraging (not sure if all can follow this link) https://docs.google.com/forms/d/1G7pi27dNkUHfYFbj94b4c5jLOiFHySG5_xqpl6bVDf8/viewanalytics If you can't see the link here is a pdf: https://www.dropbox.com/s/2uhb8ub6fjguonc/Survey%20Draft%201%20-%20Google%20Forms%20Analytics.pdf

mmnoo commented 10 years ago

I am about to start working on data export.

I am wondering which fields should be available in that and am a bit concerned about the postal code and HUD/VIN numbers.

Postal codes (at least in Canada) sometimes have so few people in them, that they are excluded from health studies because it is considered a privacy breach.

Can VIN/HUD numbers be used to identify people? Do we need to look into obscuring those numbers, or is the data relevant without including them? Also, if we include the numbers, could it make home resale hard and be a disincentive for people to give us info?

shapironick commented 10 years ago

Great questions.

This overlaps with #34 Lets not share Zip/Post codes. Definitely share State and Country. I could go either way on City.

Let's not share last four digits of VIN and last three of HUD.

Thanks thanks thanks!

mmnoo commented 10 years ago

Just a progress update, I am going to have to go over budget with data export. MongoDB is very unfriendly to table joins, more so than expected, and the fact that JavaScript is asynchronous, doesnt make it easy to do the join in the application code.

Im thinking there is a chance resolving this could take the rest of the week.

mmnoo commented 10 years ago

OK, another update on this. I have resolved the issues blocking the exporting of our data. You wont see this functionality yet on the site you have the link to, but I just wanted to let you guys know that it didnt go into crazy overtime after all, and I have moved on to working on other stuff (the welcome modal box/tour)

shapironick commented 10 years ago

Wonderful! Thanks Melissa for your swift work!

On Thu, Aug 7, 2014 at 12:16 AM, Melissa notifications@github.com wrote:

OK, another update on this. I have resolved the issues blocking the exporting of our data. You wont see this functionality yet on the site you have the link to, but I just wanted to let you guys know that it didnt go into crazy overtime after all, and I have moved on to working on other stuff (the welcome modal box/tour)

— Reply to this email directly or view it on GitHub https://github.com/publiclab/wherewebreathe/issues/14#issuecomment-51410811 .

jywarren commented 9 years ago

Removing this feature for security/privacy reasons.