An Exploration of Search Visualization Strategies 6/4/2020 1
Introduction Me: Daniel Worley Who I work for: Opensource Connections What do I do: Search Relevance Strategist 2
Overview ● Transparent Boosting (Radar) ● Histogram Facets ● High Resolution Heatmaps ● Index Statistics ● Outside of the SERP 3
Transparent Boosting ● Exposes ranking complexity in a simple format ● Allows tweaking of the ranking algorithm with no necessary knowledge of what’s under the hood 4
Radar Chart ● By plotting each boost as a scalable vector on a 2D plane we can see the impact each boost has on the ranking. ● By letting users manipulate the vectors, they gain more control over the ranking 5
Radar Chart 6
Demo http://search-viz.dev.o19s.com:81/ 7
Histogram Facets ● Provide visual indicators of a field’s cardinality ● For date ranges, this can reveal periods of higher activity which may make it easier to select a date range 8
Demo http://search-viz.dev.o19s.com:81/ 9
High Resolution Heatmaps 10
Heatmap Limitations ● Slow ● Limited to lower resolution ● Might blow up ● But... 11
What if... ● Execute multiple parallel requests and “stitch” the resulting image 12
WMS A Web Map Service (WMS) is a standard protocol developed by the Open Geospatial Consortium in 1999 for serving georeferenced map images over the Internet.[1] These images are typically produced by a map server from data provided by a GIS database - Wikipedia 13
HighRes Heatmaps ● Setup an API to provide WMS tiles ● Control over the rendering allows for multiple styles ● The client side WMS source is responsible for stitching the images 14
Demo 15
Index Statistics ● Visualizations can help you understand the data inside your index ● Transform a dump of textual information to a meaningful image 16
Avoid pages like these 17
Bearded Wookie Returns Demo 18
Thinking Outside the SERP ● Think outside of a listing of search results, maybe outside of a web browser altogether ● Unity/UE4 can drive engaging experiences powered by search 19
Outside of the SERP ● Augmented Reality search experiences ● Dynamic Museums with MLT based on what you’ve spent time looking at 20
Thank You Repos: Lukedive: https://github.com/o19s/lukedive Search-Viz: https://github.com/o19s/search-viz Heatmaps: Coming Soon 21
Recommend
More recommend