https://hoaxy.osome.iu.edu/#query=cocacola&sort=mixed&type=Twitter&lang=

Using this tool, and reading it like tea leaves, I could conclude that Sean Spicer is running an influence campaign targeted at cocacola.

I didn't have to manipulate this map at all. This is just what the tool produces. I find this, on its face, dubious to say the least.

Here we have Nike. I haven't scoped this brand's network in a long time, but this graph also looks a little...incomplete we'll say. https://hoaxy.osome.iu.edu/#query=nike&sort=mixed&type=Twitter&lang=

Again, this is just what the tool generates.

And an easy one for "owl fact" since I can easily check my own metrics for this. https://hoaxy.osome.iu.edu/#query=%22owl%20fact%22&sort=mixed&type=Twitter&lang=

This map is missing every follower that actually interacts with those tweets. It also has the founder of my agency listed as "more botlike" than me. Both of our accounts score 0% on @BotSentinel

I haven't had much time to test this tool, but what I'm seeing so far suggests it still needs a lot of work to be reliable and/or useful for researching twitter trends. There are better tools out there for this.

This thread isn't about promoting those. The important thing to take away from this is that digital tools can be highly unreliable. You have to test them, and take findings from them with a grain of salt. Otherwise, apophenia takes over and you can draw some wild conclusions.

Some quick tips on testing research tools.

1) Have a control group. Ideally something with dataset you control and can access. A comparison without a baseline is meaningless.
2) Test master brands and compare one tool's findings to other industry standards.

3) Look for documentation on the tool's methodology. Make sure it's measuring what you're trying to measure.
4) Understand the tools dimensions and metrics. Make sure the dimensions, themselves, are compatible with the paired metrics.

5) Run a test. Create a dummy hashtag or keyword from an account you control. Track the path of that media manually. Now compare that with the tool's findings. How accurate is this tool? What is it missing?

6) Take every tool with a grain of salt. They're limited. The more a tool boasts of its power, the more you should check under the hood to make sure that isn't just marketing hype.