Interactive IQR Outlier Detection and Moving Average Forecasting Tool (Upload Your Data)
IQR Segmentation Tool
Upload a CSV file containing a single column of numbers (no header, no symbols):
2‑Period Moving Average Forecast Comparison (With Outliers vs. Outliers Replaced)
How to Use This Tool
Prepare your data: create a CSV file containing one column of numbers with no header row. Upload your file and click Process. The tool performs IQR outlier detection, segmentation, and statistical profiling. It also generates two 2‑period moving average forecasts: • Scenario 1: Raw data • Scenario 2: Winsorized (outliers replaced with limits) Both forecasts appear on the same chart for comparison. The original raw dataset of values is also plotted on the chart against Scenario 1 and Scenario 2. Below is a summary of what the chart above tells you about how statistical outliers are impacting your forecasts. Simply look at how different the scale is for the Scenario 1 vs. Scenario 2 lines to see the impact. The bigger the difference between the two scales, the more outliers are having a negative impact on the accuracy of your data forecasts.
What the Chart Is Actually Showing You
1. Raw Data (grey)
This is the unprocessed series exactly as uploaded. It shows:
The true volatility of the data
Where the spikes and dips occur
Whether the series is smooth, noisy, trending, or erratic
This is your “ground truth.”
2. Raw Data MA(2) – Scenario 1 (blue)
This is a 2‑period moving average applied directly to the raw data.
It tells you:
How the short‑term trend behaves when outliers are allowed to influence it
How sensitive the forecast is to sudden jumps
Whether the series is stable enough to forecast without preprocessing
If the blue line swings sharply whenever the grey line spikes, that’s the outliers pulling the moving average around.
3. Winsorized MA(2) – Scenario 2 (red dashed)
This is the same moving average, but applied after replacing outliers with the IQR‑based limits.
It tells you:
How the trend behaves when extreme values are capped
Whether the underlying pattern is smoother than the raw data suggests
How much outliers distort short‑term forecasting
If the red line is noticeably smoother than the blue line, your outliers are materially affecting the forecast.
