> For the complete documentation index, see [llms.txt](https://measures.gitbook.io/multiple-measures/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://measures.gitbook.io/multiple-measures/progress/week-3.md).

# Week 3

## Summary to Implementation

![Interface Overview](/files/-Ll1ifMyGQoeEM1eNRtH)

* Finished: 
  * current: <https://shellywhen.github.io/measure-flow/?session=demo&datasetName=scientists>
  * framework: quite-independent modules including node-link view, configuration bar, time slider, multiple-layer timeline view (for global measures), connected scatter plot view
    * Issue:&#x20;
* To do
  * Interaction
    * 😵 flexible selections - support search and add new node/link selections
    * 🔥 encode more features&#x20;
      * line: link type, link strength
      * circle: local measures in a brushed period (on demand)
      * 🔥 measures: local / global
      * highlight: outliers in the value series
  * Complementary Views
    * 1-d small multiples for connected scatter plot
  * Embed more measures:
    * link-centered: centrality
  * 🔥 More data
* Problems
  * selection is quite tricky and the source code is hard for me to understands
  * performance improvement: unclear of data storage and webGL, implementation of measure calculation is ineffective, should use matrix to obtain values instead of iteration over links.

{% hint style="info" %}
**Note**: time aggregation is based on a greedy approach: given a granularity like day, do the aggregation from the start point.
{% endhint %}

## Ideas for Discussion

### Measures in Different Layers

"aggregated" refers to measures with a single value in a static network (dynamic network aggregated by time, i.e. network obtained using the brush) or the derived value from a series of networks.

"series" refers to a series of values obtained from a series of networks.

* local
  * aggregated: encode values on node-link diagrams
  * series: connected scatter plot for comparison or simple 1d line chart.
* **subgraph**

  ⭐ **Create the other aggregation of the graph, whose node represents a subset of initial nodes. Use a popover to handle complicated interactions. Normally, a subgraph is greatly smaller than the original graph, with several groups taken into consideration.**

  * aggregated: new node-link diagram
  * series: a multi-label timeline list for different measures
* network
  * aggregated: a list of values or a more flexible calculator (?)
    * **⭐ local measure distribution over nodes**
  * series: multi-layer timeline view as James'.

### Design Choice

see above to do list