"Finding The Pigs" shares how quickly Python can be used by beginners to analyze social media data to better understand an audience segment.
Code for this activity can be found at https://github.com/trufflemedia/find-the-pigs/
15. Code looks for terms in the Twitter
name and the Twitter description.
16. But with the data above, I feel confidant
@Swineweb’s followers do not represent 10,000
pork producers or veterinarians.
And the followers of the other swine media Twitter
names don’t have 10,000 pork producers or
veterinarians either.
Did I find the pigs?
First, Twitter information is self declared. It is
possible there are many more pork producers
following Swineweb’s Twitter account. But they are
not telling the world of their pork or vet intentions.
17. And the followers of the other swine media Twitter
names don’t have 10,000 pork producers or
veterinarians either.
Did I find the pigs?
First, Twitter information is self declared. It is
possible there are many more pork producers
following Swineweb’s Twitter account. But they are
not telling the world of their pork or vet intentions.
But with the data above, I feel confidant
@Swineweb’s followers do not represent 10,000
pork producers or veterinarians.
18. No John, we just got to 10,000
followers and thought we would
share that number as a goal
reached.
Hi Jim, am I missing anything
in your “10,000 pork
producers following you”
statement?
19. Python can be
picked up quickly
There are lots of
resources online
What did I learn?
There are other Python analysis
tools that could offer more insight
Editor's Notes
Hi my name is John Blue with Truffle Media Networks
My lightning talk is “Finding The Pigs”
We produce a podcast for top tier pork producers, swine veterinarians, and animal health researchers call SwineCast.
The pork industry is a small segment of agriculture.
And there are only few media properties that cover the business of pork.
In August Swineweb, one of our competitors, tweeted they “reached record 10,000 pork producers, veterinarians on Twitter.”
And their press release expanded “More than 10,000 pork producers and industry professionals are following Swineweb’s Twitter feed to receive their daily news and market updates.”
I was skeptical of the claim.
I looked at the follower list of @Swineweb. With a quick review I could see that there were several followers that had no indication of being involved in pork production, being a veterinarian, or involved with farming in anyway!
To learn more and make sure I was not missing something, I used Audiense to review the @Swinewbcom Twitter followers. Audience is a service that offers quick and easy access to the Twitter data.
The followers on this first page were not involved in anyway with agriculture. Here are a few examples.
Zach Boychuk, looking for Instagram and Snapchat connections.
And Ziplock, promoting his rap and musicality.
I still was not 100% sure. I didn’t have any scaled data to really support me saying their claim was bogus.
Is there a way to programmatically review all 10,000 @Swineweb followers to get some sense of who they might be or what interests they have?
Possibly. This is where Python, the programming language, comes in.
Python is not my expertise, my rank = beginner.
I had taken some lessons from an on line course called “Biology Meets Programming: Bioinformatics for Beginners”, which used Python to look at DNA sequences to extract patterns.
Since I was looking for patterns in a spreadsheet, I Googled python and spreadsheets: Yes there were utilities to read & write to them.
I thought if I could read some info from a spreadsheet, I could scan it to discover more about the followers of @Swineweb.
I ran with that info to create a program to scan Twitter follower info.
I was interested in the Twitter name and description. Using Audiense, I exported the Swineweb Twitter followers as a spreadsheet. I planned to scan each follower’s name and description for a set of about 50 terms and words I would give the program. And I planned to end up with another spreadsheet of the Twitter names and a 1 for each term found.
On the right are the terms.
Here is a rough outline of the Python code action: For each follower look for specific terms and output a 1 (yes, term exists in the name + description) or 0 (no, term is not in the name + description).
Did I find the pigs?
First, Twitter information is self declared. It is possible there are many more pork producers following Swineweb’s Twitter account. But they are not telling the world of their pork or vet intentions. (Next)
But with the data above, I feel confidant @Swineweb’s followers do not represent 10,000 pork producers or veterinarians.
And the followers of the other swine media Twitter names don’t have 10,000 pork producers or veterinarians either.
Did I miss anything? Was there market research? Did @Swineweb have some other source of data that offered better detail?
I called up Jim and Swineweb: Hi Jim, am I missing anything in your “10,000 pork producers following you” statement?
No John, we just got to 10,000 followers and thought we would share that number as a goal reached.
What did I learn? Python can be picked up quickly; There are lots of resources online; There are other Python analysis tools that could offer more insight.
And when it comes to social media marketing, there are a lot of smoke and mirrors.
Remember: If you’re not paying for it, you’re not the customer.
You’re the product being sold.