Advancing the Art of Big Data Analytics
The popular Netflix show ‘House of Cards’ has been all the rage amongst viewers. In fact, viewers were so enamored by it that many watched the entire second season in one go. It’s added to our lexicon – Binge-viewing! How do you think Netflix got the idea for this blockbuster show? “Big Data” analyses allowed Netflix to pose the question – statistics such as streaming and rental habits and preferences from various audience segments were used to identify the opportunity and discover the blockbuster solution – an American version of ‘House of Cards’. The genius in this was the ability to pinpoint the perceived need of the audience based on viewing preferences and then realize a concept to address it. Another well-concealed achievement here was the ability to predict the success of the show (and hence invest in one). Herein, lies the biggest utility of Big Data – the ability to unearth hidden relationships among a diverse set of data and then use it in a suitable application.
Big Data is very hot and understandably Silicon Valley has happily lapped it up. Try eavesdropping on a coffee conversation in Palo Alto and I am sure you will have heard enough and more on Big Data. So what do people who work with Big Data actually do? Some of us in the field make the whole data mining process sound ridiculously easy- in simple words an algorithm when applied, opens a treasure chest of money making possibilities. If only it were that easy! Big Data is much more than storing, organizing and crunching all the log files you used to throw away. Defining Big Data as what can be done with using technical jargons like Hadoop or NoSQL amounts to a hijacking of the term. Winston Churchill once famously said that, “True genius resides in the capacity for evaluation of uncertain, hazardous, and conflicting information.” He was absolutely spot on. Comprehending every aspect of the data is critical to success. Data analytics is an art, almost akin to cooking a meal with the Chef being the Chief Analyst. One may be equipped with all the latest technology tools, in our Chef’s case the cutlery and ingredients, yet the difference between good to great lies in the algorithm or casino online the hand that uses the knife to slice and dice, judiciously mixing ingredients resulting in that yummy dish.
So, where is the promise of Big Data in pharmaceutical applications? The key to using Big Data is to utilize all of the wonderful advances in data sciences and to ask the right questions (and hence obtain the relevant answers) that were previously either impossible or time consuming. Consider the following – to understand the key elements in conquering a major disease, one has to integrate epidemiology, clinical data, patient outcomes data, molecular pathways, pharmacology, drug delivery, translation in the clinic etc. etc. Until now, each of these aspects had a specified way of classifying data and in order to integrate store and retrieve knowledge from all these fields one had to have armies of domain experts– often incompatible with each other. For example, try integrating patient outcomes data with genomic and pharmacological data using classic relational database approaches! One would have to recreate an entire new database with complex relationships defining how one can query the database. This led to significant inertia in using advances in each of these fields for a common goal – curing the disease!
Advances in Big Data technology has allowed us to “acquire” related data from each of these diverse domains and then integrate the pieces of information so that those experienced in the art can ask the right questions to obtain relevant answers. PharmGPS® uses a Big Data approach to integrate these diverse domains and present mined knowledge to experts so that they may utilize this knowledge in devising the best way to tackle a disease. This may involve mining information and data in different formats such as DNA sequence, patient records, chemistry structures, clinical trial data, molecular pathways – all to understand how to correct, reverse or manage the pathophysiology in any given disease. Finally, all this needs to be presented in a friendly and intuitive user interface that simplifies usage but does not hide complexity. PharmGPS® was designed and implemented to attain the above goal and is being perfected with every additional collaboration within the Pharma industry.
On a lighter or culinary vein – here is how PharmGPS® plays the expert Chef in helping translate big data to “knowledge” for “enabling decisions”.
Deep-fry 2 cups of the Unstructured Data (EMR, Scientific, Clinical, Epidemiological, Commercial). Integrate and unify to strategizeonwhere you should play in the future. Keep stirring for 2-3 minutes. Reduce Heat. Add 4 tbsp of the PharmGPS® proprietary algorithm to get your ‘Differentiated Target Product Profile’ and spruce up your “where to play” opportunity. Season it with a dash of ‘Performance Benchmarking’ by amalgamating internal and external factors – 450gms. Cool for 10 mins. Add 2 pinches of ‘Competitor Evaluation’ and mix thoroughly. Pour half cup of your Potential Target Licensing picks over the preparation. Wait till it get’s soaked. Garnish the dish with 2 tbsp of your ‘Go-to-Market Strategy’ and crack deals with the right evaluation before serving.