In our last couple of posts we reviewed some of the ways in which ideas from the field of data visualization can be used to reimagine the world of advanced search. For example, we’ve explored how visual representations can provide a more direct mapping between the underlying semantics of a search and its physical appearance. Likewise, we’ve shown how automated query suggestions can be used to facilitate the task of keyword generation. And we’ve shown how field tags can offer a small but significant step toward the prospect of universal representation for search strategies.
These developments offer unprecedented opportunities for the creation, sharing and optimization of novel solutions to complex search tasks. But not all search tasks are new. In fact, many search problems are variations on an existing theme, and in these instances it makes more sense to start from a previous solution. In sourcing, for example, recruitment professionals draw on repositories such as the Boolean Search Strings Repository and the Boolean String Bank as a source of inspiration and solutions to complex search problems.
However, despite the undoubted value of such resources, their content remains stored as raw text. This means that not only does it suffer from a number of inherent shortcomings, but the true value of this content as source of inspiration, experimentation and learning may never be fully realized. Which is why we’re pleased to announce this week the release of support for Boolean string parsing. This means you can take (almost) any Boolean string and have it rendered in a form that it is amenable to interactive exploration and experimentation.
For example, here’s one of the strings posted to the Boolean Search Strings Repository, for a Network Engineer:
(network OR networking OR networks OR internet OR ip OR ipv4 OR ipv6 OR “internet protocol” OR iptv OR vpls OR ethernet OR tru2way OR “Cisco Certified Internetwork Expert” OR ccie) (engineer OR architect OR designer OR design OR engineering OR architecture OR developer OR develop OR developed OR expert OR sme OR specialist OR guru OR ninja OR analyst OR analysis OR consultant OR “Cisco Certified Internetwork Expert” OR ccie) (ip OR “internet protocol” OR ipv4 OR ipv6 OR broadband OR ethernet OR cat5 OR “triple play” OR “quad play” OR virgin OR “bt vision” OR “talk talk” OR iptv OR vpls OR mpls OR tdm OR eompls OR ott OR isp OR “adaptive streaming” OR tru2way OR openway) (development OR developing OR developed OR built OR building OR build OR constructed OR construction OR construct OR implemented OR implement OR implements OR design OR designed OR designs) (cdn OR “content delivery network” OR “cdns” OR “content delivery networks” OR “content delivery networking” OR p2p OR “p-2-p” OR icap OR “opes” OR esi) (telco OR telecoms OR telecommunications OR telcos OR o2 OR virgin OR vodafone OR “everything everywhere” OR bt OR orange OR tmobile OR “talk talk” OR cdn OR “triple play” OR “quad play”)
In this form it looks pretty impenetrable, but when we enter it into 2Dsearch we start to see the wood for the trees:
As we can see, it’s a conjunction (AND) of 6 disjunctions (ORs). So now we understand its structure, we can optimize it further, e.g. by:
- Experimenting with search suggestions to refine each of the disjunctions (OR groups)
- Naming the OR groups and save them individually (like lego blocks we can re-use).
In each case we would of course keep a close eye on the search results pane, noting how they update dynamically with each change.
Here’s another example, this time for a data migration project manager in Dublin:
(cv OR “cirriculum vitae” OR resume OR “résumé”) (filetype:doc OR filetype:pdf OR filetype:txt) (inurl:profile OR inurl:cv OR inurl:resume OR initile:profile OR intitle:cv OR initile:resume) (“project manager” OR “it project manager” OR “program* manager” OR “data migration manager” OR “data migration project manager”) (leinster OR munster OR ulster OR connaught OR dublin) -template -sample -example -tutorial -builder -”writing tips” -apply -advert -consultancy
And now in its visual form:
So again, a conjunction of OR clauses (the pale blue blocks), but this time with a sprinkling of field tags (dark blue) and various negated terms (white on black). Moreover, in this example we can see that the search results don’t appear to be particularly relevant, so we might start to experiment and optimize by:
- fixing the typos in the keywords and the field tags
- enabling/disabling the location group (to understand what effect this clause has on the overall expression, i.e. whether there any relevant results out there at all regardless of location)
- experimenting with different location keywords using automated query suggestions
- splitting the skills keywords into two groups, e.g. one for the job role and another for the technical skills
And so on.
Hopefully the above examples communicate the value of data visualization techniques in facilitating the understanding and exploration of search strategies and Boolean strings. Over the coming weeks we’ll be keeping an eye on the communities mentioned above and sharing any noteworthy examples here, with our own insight and commentary. However, if we haven’t convinced you, and Boolean strings remain your thing, that’s fine — you can still use 2Dsearch to experiment and refine, then simply export your work as a traditional Boolean string (just hit the ‘copy’ button on the Query tab).
In the meantime, if you have any scary strings of your own to share, send them in and we’ll see if we can help make sense of them. Or just head on over to 2Dsearch and try them out yourself!
Originally published at www.2dsearch.com.