Hi ,
How do you extract trustworthy insights from data you know has been deliberately manipulated?
It's a challenge most data scientists never face.
We're used to cleaning messy data, but deliberately manipulated data? That's completely next level.
Yet if you're working with social media data, manipulation is the reality you're dealing with.
Tim O'Hearn, a reformed social media hacker who generated millions of followers through bot manipulation, recently shared with me the harsh reality:
"During what I would describe as the golden age of Instagram botting, (the proportion of fake accounts) was probably as high as 40%."
Let that sink in for a moment.
If you're making business decisions based on social media data, nearly half of what you're analysing could be artificial bot activity.
And if you're attributing
value to social media accounts without filtering for bots, you're potentially wasting 10-20% of your marketing budget on fake audiences.
The good news is there are ways to identify and filter out this manipulated data - techniques that can also apply to identifying suspicious records in any dataset.
In the latest episode of Value Driven Data Science, Tim joins me again to reveal practical strategies for identifying and filtering out bot activity from social media datasets to extract trustworthy business insights.
This Value Boost episode uncovers:
- The telltale patterns in social
media data that reveal bot activity [03:10]
- How machine learning classifiers can identify bot accounts [05:20]
- Why removing bot activity can increase marketing ROI by 10-20% [06:41]
- The broader application of these techniques beyond social media for identifying "dodgy" data records in any dataset [07:25]
Essential listening for anyone working with social media data.
🎧 Listen now on Apple Podcasts or Spotify, or click the link below:
Episode 73: How to Trust Social Media Data When You Can't Trust Social Media
Talk again soon,
Dr Genevieve Hayes.
p.s. Next month, I'm teaching 3-5 data scientists my complete process for creating your own high-value data science opportunities in the Data Science Impact Sprint - a 4-week, 1-on-1 coaching program that will boost your strategic influence and help position you for career advancement.
Reply with "SPRINT" and I'll send you the details.
Doors close at 9am on Saturday 2nd August Melbourne, Australia Time (7pm Friday 1st August US EDT) or when all the places fill.
