Three Easy Ways to Eliminate Noisy Data

Three aspects of this tip are to avoid data inserted by others, use smoothing techniques like rolling averages, and avoid throwing out the outliers.

Decision makers can work with imperfect data. What they won’t tolerate is manipulated data. Avoid shortcuts that change the picture, even if they look convenient. Present the data as collected, explain the limitations, and keep your credibility intact. The tip is simple: be true to the data.

 

It Starts with Consistent Data Collection

Ensure your data collection methods are consistent and accurate. This includes calibrating measurement instruments and standardizing data entry processes to minimize errors.

Three Easy Ways to Eliminate Noisy Data

1. Inspect for and Avoid Third-Party Data Cleaning

The enemy here is imputation methods to fill in missing data. Often, someone in the data chain will attempt to fill gaps before handing over the collected data to you.

2. Use Smoothing Techniques

Use moving averages. This method calculates the average of a fixed number of data points to smooth out short-term fluctuations and highlight longer-term trends. It's simple and effective for time series data.

3. Avoid Removing Outliers

Honesty is the best policy when dealing with high-level decision makers. Remember, you get one shot at credibility. Don’t be perceived as doing something that paints a picture at odds with the collected data, even if you think filling gaps or removing outliers tells a better story.

The Tip: Be True to the Data

Three aspects of this tip are to avoid data inserted by others, use smoothing techniques like rolling averages, and avoid throwing out the outliers. The tip for effective communication is to be true to the data.

---

The elements of the FINESSE Fishbone Diagram® are Frame, Illustrate, Noise reduction, Empathy, Structure, Synergy, and Ethics.