Unlike, most of the other speakers I am not a GIS professional.
Instead, my talk focused on the value, and the challenges, in creating a link between spatial data – typically used by the GIS team – and relational data – used by IT and most business users.
For my theme I chose “Meeting your customers where they live”
A key use of big data analytics is to analyse customer behavior and identify which interactions lead to purchases.My premise is simple – to improve our customer’s experience we need to offer him, or her, the option to deal with us at the time and place of choice.
If we can translate the address information we hold about our customer into GPS coordinates (geocoding) and compare these coordinates to where our business has a physical presence we can increase the chance that our customers will use for service, rather than choosing a more convenient competitor.
This can also help with planning – if we can identify where our target market does not live – through an analysis of location data – we can avoid investing in infrastructure where it is not needed. One telecommunications customer of our partner, Datameer, saved hundreds of millions of dollars by avoiding upgrading their infrastructure in low demand areas.
However, in order to this the client had to first improve the quality of location data available to them.
Missing or inaccurate address data creates ambiguity that can make it very difficult to accurately identify a unique address and assign a geocode.
Most of us are familiar with Google Maps.Google Maps identifies missing, or inaccurate information and prompts us to resolve the ambiguity. If, for example, we type “Smth Street, Johannesburg” into Google search it will ask if we meant “Smith Street”, or “Smit Street” and allow us to improve our search.
To automatically enhance our corporate address data, running into millions or tens of millions of address records we need to reduce this ambiguity in order to get reasonable results. We must correct obvious errors and, where possible, add missing information. In South Africa, we must standardise on a one variation of a place name to simplify matching to a suitable reference data set.
I call this process “Bridging the Gap.” [Tweet this]
Some of the complexities of improving South African address data quality were covered in my recent post.
Without quality location data we cannot hope to meet our customers where they live.
How does the quality of your data affect your customer’s experience?
image sourced from https://commons.wikimedia.org/wiki/File:PSU_no_load_5_star_rating_chart.png