Trendspotting: Choose C-Phrase!
Fred works at an insurance company and has noticed that payouts to homeowners have been trending up over inflation starting about two years ago. He would like to raise rates, but he must provide his regulator with specifics. Who should he raise rates for? Who is causing this increase and why? When Fred looks at the data all he sees is a noisy and sparse graph of payouts over time. How is he going to make sense of this?
Luckily for Fred, his company has recently installed the C-Phrase plug-in over their database of payouts. Fred types the question, " total quarterly payouts on homeowners policies for the last 36 months " . It shows they have trended up starting about 8 quarters ago. Fred continues, " group by cause of loss category ". Now he sees this broken down by the four main causes: Fire , Wind , Water , Human . Wind and human losses are essentially flat, but fire and especially water losses are trending up. For now he explores the water cause. After several more detailed interactions, Fred has produced a graph that clearly shows that for houses in a specific sub-region of the country there is a sizable increase in Appliance Malfunction Water Damage . After calling some field agents and investigating further, Fred concludes that the cause is a specific water boiler model that was sold into that market 20 years ago. Fred will prepare an argument for a targeted rate increase, but now he goes back to C-Phrase delighted to explore more aspects of the data. Fred marvels, "finally a tool that lets me see the forest for the trees!"
Many variants of the problem Fred faces arise in mature businesses (e.g. insurance, utilities) that must spot trends in sparse or noisy data. C-Phrase gives solid benefits for analysts to quickly and easily explore such data in an infinite number of ways.
There are many tools (such as Tableau) which let analysts visualize data. These tools however requiring programming support to set-up highly conditioned trend analysis, such as " show the payouts of Water damage per 100000 homes for non-apartment homes grouped by regions. " To set up such a tableau report requires expertise and time. In C-Phrase you just type the command in. Some may say that AI or data-mining can solve Fred's problems. We don't think so -- at least not in the short term. Fred knows his market and was following his hunches as he explored. Fred can explain his reasoning. AI generally cannot. C-Phrase faithfully assisted Fred to explore and back up his hunches with hard evidence. Fred using C-Phrase will beat all the other alternatives for some time to come ... possibly forever.