Use the Correlations panel to discover the relationships between variables in your data. After you build any kind of Graphext project, you'll be able to generate correlation charts that reveal how the values belonging to one variable are associated with the values belonging to another.
“Correlation doesn't imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mounting 'look over there'.”
- Randall Munroe
Charts in your Correlations panel reveal the density of your dataset at points where values from two variables meet. Your y-axis represents values from the variable in your search bar and the x-axis represents values from the correlated variable - labelled in the top right of each card.
The blue circles in your correlation charts represent the density of data at each value intersection. Bigger and brighter circles represent a higher number of data points at an intersection whereas lower and duller circles represent fewer data points at an intersection.
A strong positive correlation would be signified by a trend of big & bright circles moving diagonally upwards from left to right 📈
A strong negative correlation would be signified by a trend of big & bright circles moving diagonally upwards from right to left 📉
You can determine the strength of correlation in Graphext Correlation charts by examining the statistic in the top right of each card representing the relevance of the variable. Chart cards with higher relevance statistics and more white bars represent a stronger correlation.
Correlation charts are ordered in terms of their relevance. Relevance scores refer to the mutual information shared by the two variables rather than linear correlation. Mutual information is more powerful than linear correlation in detecting arbitrary associations between variables.
Across Graphext's UI, relative mode presents data as a proportional representation. In practise, this means you see data as a percentage distribution rather than an absolute count.
This is especially useful in Correlations charts because these use size and color to visualize the correlation between lots of values belonging to pairs of variables. With relative mode, the size and color range of bubbles in Correlations charts are restricted to a percentage distribution (either on the x or y axis). This makes it easier to spot patterns.
To start plotting correlation charts, navigate to the Correlation panel inside a Graphext project. Choose a variable using the search box and start exploring correlations between variables.
Use the search bar at the top of your Compare panel to change the core variable that you are correlating with other variables in your data.
Saving your Correlations charts as insights allows you to enhance your chart with elements like titles, descriptions and statistics. You can then present your Correlations charts directly inside of your project's Insights panel.
Exporting Correlations charts individually means you can quickly include them in reports or presentations. To export a Correlations chart, use the 3 dots on the top right of the chart's card. Customise the appearance, size and theme of your Correlations chart in the download window.
Correlations charts are exported individually. To export more than one, first create an insight featuring the charts you want to export, then export that insight.
Correlations charts are generated for lots of variables in your data. It's likely that not all of these will be useful so hide the ones that aren't. To hide a chart from your project's Correlations panel, start by clicking on the three dots at the top right of the chart's card.
Hiding charts can be useful when you are capturing an insight from the first 5 Correlations charts.
We know that data isn't always clean and simple.
Have a look through these topics if you can't see what you are looking for.