In addition to helping you understand complex relationships, regression allows you to make specific predictions.

You can do that two ways:

1. Entering numbers manually where you see the words "Enter a value".

For example we could enter in the properties of a house and this model will predict the value of that house: For example if we enter in the values...

• Sqft Tax Record = 1700
• Close Date = 1850
• Total Baths = 2
• Central Cooling = No

...then regression will spit out the prediction of a Sold Price of \$259,896.

When you hover over the prediction, Statwing also provides a confidence interval around the prediction to give you a sense of how accurate the it is likely to be: Dates are a bit tricky here. Dates can't be directly entered into a regression, so Statwing automatically changes them to be "Number of days since the first day in the sample." That is, if your data goes back one year, the days will be changed to 0 through 365. So, if you want to predict a value using a specific date, you'll have to calculate how many days that date has been from the beginning of the dataset.

For example, if the dataset goes back five years, and you want to predict "Sold Price" right now for a property, you'd enter in the most recent date, which is "825" (because 5 * 365 = 1825). A quick way to find this for any dataset is to hover over the end of the relevant histogram and it will tell you the last number available (below, 1850). 2. Getting predictions for all data in the dataset

Statwing also provides a "Make Predictions" button: Clicking that will add another column to the end of your data, where Statwing will automatically do what we did manually above, making a prediction for every value in the dataset (whether or not that data already had a "Sold Price" value when uploaded).