| In this Session... | Before you begin... |
|---|---|
| * Using Filters to display and animate time series., Using the f(x) Transform to reformat data., Using Parametric and Geometric layouts and the Quick Layout menu with Filters. | To follow along, you can download: HowTo_09START_2_14.graph... and the snapshots library: HowTo_09START_2_14.graph... |
| Slide | |
|---|---|
| 1 | How To GraphXR 9. Time Series |
| 2 | Before You Begin... Ideally, you'll have worked through Module 8. Layouts and Pinning. If you're starting here, and you want to follow along, you'll need to: * Log in to GraphXR, create a Project, and open its graph space., Download starting data (HowTo_09_START.graphxr and the snapshot library HowTo_09_START.graphxrsnapshots) for this module and drag and drop it onto the graph space. |
| 3 | Open a Table to see that our Episodes nodes include an episodeAirDate property. To visualize time series in the graph we can use parametric or geometric layouts to order nodes by date-time property values. We can also filter nodes by a date or date range. |
| 4 | First, we'll go to Project=>Category and click Episodes. We'll set episodeAirDate values as captions and scale the nodes by the millionViewers property value. |
| 5 | Open a Table to review episodeAirDate property values. GraphXR's preferred format for a date and time stamp is YYYY-MM-DD HH:mm:ss. We can edit incorrect data formats either manually or using a preset Transform. |
| 6 | If just one or a few dates are in the wrong format, you can simply edit the property value, either in the node's Information window, or an Enhanced Table. |
| 7 | If most or all dates need reformatting, you can use a preset formula in Transform=>f(x). In the Transform panel => f(x) tab select the category (Episodes) and property (e.g. episodeAirDate) that has the date-time data you need to transform. |
| 8 | Select the toDate or convertCustomDate Input Formula. Enter a New Property Name for the converted data. A sample result appears below. Then click Test to test the transformation or Run to perform it. |
| 9 | Now open the Layout=> Parametric tab. Select episodeAirDate and millionViewers for X and Y axes and totalCharacters for the Z axis, then click Apply. Next we can apply a filter to isolate nodes with specific date property values. |
| 10 | Open the Filter panel. To filter Episode nodes by air dates (or range of air dates), select episodeAirDate in the Node Property menu. |
| 11 | You can manually scrub through the values using the handles, or use the playback controller to play through values (or a locked range). This lets us animate filtered data and share the animation with other users as a view. |
| 12 | Filters also provide effective visualization when used with the geometric or geospatial layouts that we can apply to different categories of nodes. |
| 13 | In the Geometric tab click Order by Property, select Episodes and the episodeAirDate property, click Ascend, and click Grid. You can then use Expand on the Quick Layout toolbar menu to expand the grid. |
| 14 | Now repeat the process to lay out Character nodes in a spiral, and House nodes in a line. Then, in the Legend, select both categories using ctrl+left click, and move those nodes off the grid using left click+ drag. |
| 15 | Now open the Filter panel and again select episodeAirDate in the Node Property menu to filter Episode nodes by date. Notice that we can apply multiple filters for any numerical or date property (e.g. the kills property of Character nodes). |
| 16 | Let's filter on a range of a 2-month period. Use the handles to set a range, or click the the Min or Max fields to select a precise date and time. Select Lock Range to filter data in 2-month intervals. |
| 17 | Again, scrub through the values using the handles, or use the playback controller to play through values (or a locked range). |
| 18 | Click the trash can icon on the filters to clear them and display all the data again. |
| 19 | Next, in Module 10. Algorithms, we'll look at how to measure graph patterns in terms of connectedness, paths, and grouping of nodes. |