I keep meaning to document the data practices for my StickersAndStamps project. As the project keeps growing, I get behind in writing up my evolving process. Please keep in mind that this work-flow has evolved over the years that I’ve been working on this project.

What happens when you sign up:

1. I use a google form to provide a private way for folks to sign up and send me their postal mailing address, so that I can send sticker and note to them through the US Postal Service.

2. The form entries are automatically saved to a private google spreadsheet, as a function of google forms.

What happens after you sign up:

1. When there are several new signups (20+), I copy the new form info to another private google spreadsheet, my ‘main’ data set. Then I download the updated spreadsheet as a CSV text file.

2. I’ve written python scripts that I run on my laptop to process the downloaded CSV file.

3. One of these scripts takes the downloaded CSV file and generates an easy-for-me-to-read list of mailing addresses. I include the sticker choice, if they want one of my temp tattoos, and comments, so that I know what to include with the note. 

4. print this list on paper, and use it when I hand-write notes and address envelopes for the folks who have signed up. The hand-writing is important…it is a meditation for me, and a special treat for the recipient. We so rarely receive hand written/addressed mail.

5. Then I shred the printed pages of addresses, and recycle the paper shreds.

What then? How do I use your email address?

1. When I’ve gotten the letters ready, I send out emails to let folks know their letter has been mailed and to start checking their mailbox. I walk to my local post office and put the envelopes in one of the blue boxes out front. I save the date I mail those envelopes in my main data set (in a google doc).
2. When folks reply to my email, I use this email as a search term in my growing google-sheet to search for the request and save the arrival date in my main data set. Fairly often, people reply with a completely different email address than they signed up with. In these cases, I have to try searching for their name or other info in their email. Many times, there is no way for me to guess which request is associated with the report. In these cases, unless the person responds helpfully, to my request for clarification, there is only lost effort…I can’t use the report in my data set…because I don’t know the city,state,zip and the date I mailed it.

What do I use for analysis and visualization?

1. From the information folks sign up with, the data I use for analysis is: City, State, Zipcode, timestamp of form entry, and sticker design choice.
2. In the ‘main’ spreadsheet, I add some additional information for each request, including: lat-lon of the City, and the USPS Sorting Center associated with the zip code, the date I mailed the letter, and the date the letter was reported arrived. If the recipient sent a letter back to me, I also include the dates of sending and arriving.
3. With one of the python scripts I mentioned above, I process raw data on my laptop to generate the public-facing data set…this public-facing data set does not include personally identifying information. Zipcode is the most personal information included.
4. At this point, I have a CSV with information on the letters I’ve mailed out. With the lat-lon information, I can plot data on maps as well as charts.

-----------------------------------------------------------------------------------------------

This post was originally on Medium. But now that itch.io is supporting blog posts, I'm moving it here because the project is here.  Super happy about itch.io supporting blog posts now. Wow! Really fantastic news!