Monica Rogati, a seasoned data scientist at LinkedIn, outlined ten common mistakes that professional data analysts often make.
1. Assuming Clean and Unbiased Data:
Always scrutinise your data. It often contains biases or errors that, if unaddressed, can skew your entire analysis.
2. Not Normalising Data:
Different scales can mislead comparisons and consider the nature of the population the data comes from. Normalising helps ensure that you’re comparing apples to apples, not apples to oranges.
3. Excluding Outliers Too Quickly:
Sometimes, what appears as an outlier may hold crucial insights about anomalies or new insights. Investigate the nature of outliers before you decide to exclude them.
4. Including All Outliers:
Conversely, not all outliers are meaningful. Evaluate their impact on your analysis to decide whether they should be included. Don’t build a product or service just around outliers as they are unlikely to reflect the needs of the mass market.
5. Ignoring Seasonality:
Trends can vary by time of year, and failing to account for this can obscure underlying patterns.
6. Overlooking Base Sizes:
When reporting on growth or changes, always consider the base sizes. User growth of 100% might sound impressive, but if based upon an increase from 10 to 20 users, this may not be sufficient to sustain a business.
7. Information Overload in Dashboards:
More data isn’t always better. Clutter can distract from key insights. Aim for clarity and simplicity.
8. Over-Sensitive Alerts:
Setting alerts too sensitive can lead to frequent false alarms, causing important alerts to be overlooked.
9. “Not Collected Here” Bias:
Don’t dismiss external data sources. They can provide critical insights and a fresh perspective that your data might not reveal.
10. Focussing on Vanity Metrics:
It’s easy to get distracted by surface-level metrics that look impressive but don’t tie back to business objectives. Focus on metrics that matter.
Remember, the goal of data analysis isn’t just to collect data but to derive actionable insights that drive informed decisions. By sidestepping these common errors, you can ensure your analysis is both robust and meaningful.
💡 Eager to dive deeper into data science and analytics? Connect with me for more insights and strategies tailored to your business needs.
Finally, you can check out more of our blogs here.
#DataScience #GA4 #GoogleAnalytics #ConversionUplift
