Time-Series Data : Exploring cross-relations with synced scatter-plot view

This page is part of Exploring Time-Series Data. Examples use the world development indicators dataset.

When your data has multiple attributes measured at each time-point, such as GDP and life expectancy per year, you may wonder how two or more attributes relate to each other from a global perspective? Is there any correlation between two variables, and does the correlation change over time? Are countries getting richer and increasing life expectancy at the same time, or are there countries getting wealthier without improvement in health, or healthier without improvement in economic outcomes? Using synched scatter-plot is a good way to quickly see such trends.

Scatter-plot view in Keshif works with time-series data to help you explore these relations over time. Keshif automatically keeps the time in-sync across the scatter-plot chart.

Opening Scatter-plot View

When your data includes multiple numeric or time-series attributes, Keshif shows the scatter-plot view among view options on top-right of the record display. Click the scatter icon to open scatter-plot view. Each record in your data will be plotted as a dot, with its position based on the compared attributes.

Highlighting Records

In this view, you can select individual records to see the record title and values on both axis (selected attributes). You can click on a single record to see all the details relating to that record (such as country).

Changing Compared Attributes

You can change the compared attributes by clicking on the attribute name in either axis, which will display a drop-down menu with potential options.

Synced Exploration over Time

The scatter-plot view shows relations across two variables on a synced time key (such as a selected year).

To see changes over time, simply change the time-key of one of the variables, and all related time-series attributes will stay in sync and Keshif will animate changes to reveal your data over time. You can change the key using drop-down menu, or the slider.

Hint: You can even let Keshif animate changes over time with an automated animation, advancing time variable every 5 seconds.

Enriching your analysis by color using third attribute

You can add more depth to your analysis by changing the dot color using the color setting options.

You can change the color theme by either using an inverse-mapping (light>>dark , or dark>>light), or using two-color (bi-variate) theme which will highlight the minimum and maximum range using different colors.

When you are using an attribute for color, mouse-over on a dot (record) will show the value used for color. If the attribute is another time-series variable, Keshif will automatically plot a small time-series trend.