Evolution of the Data Scientist - From Data Drudgery to Data Science for All

MODERN DATA SCIENCE IS IN. DATA DRUDGERY IS OUT. The Right Tools and Practices Make the Difference Today State of the art data science technology is changing fast. In just a few years, analytics and data preparation tools have totally transformed. There's a visible divide between those who've adopted the latest in big data discovery and analytics vs. data science teams using old school Bl technology and spreadsheets. Below is a comparison of NOW vs. THEN. Modern Data Science for All INUW The whole company benefits from granular, on-demand insights from big data. Most of my day is spent in the front office doing data discovery to find hidden patterns that solve business problems. I work with a team of people across multiple departments. I'm busy cracking important business problems for the C-suite by analyzing data via a big data discovery platform. It takes me minutes to prepare terabytes or petabytes of unstructured data and visualize it. П I use Spark's advanced analytics capabilities to dive deeper into data in Hadoop. Where Does the Time Go? 10% Data Preparation 20% Making Decisions 20% Planning 25% Modeling & Iteration 25% Analysis & Insights THEN THEN Data Drudgery Are antiquated analytics technologies holding your company back? WAITING FOR INSIGHTS Most of my day was spent in the back of the office, preparing and cleaning data. I worked alone, siloed from other departments. I was tasked with project backlogs, hiring more data analysts, and extracting stale insights based on old data. PROCESSING 10101010101010 RETRIEVING DATA 010101 DATA 01010 110 00 It used to take me months to prepare gigabytes of structured data within a relational database and display results in a graph. I worked with technologies such as Excel, SAS, SQL, MapReduce, MATLAB, HIVE, PIG and other BI tools. Where Did the Time Go? 10% Analysis & Insights 10% Modeling & Iteration 5% Making Decisions 10% Planning 65% Data Preparation HOW THE PROCESS HAS CHANGED NOW Streamlined, fast and powerful DATA DATA -> ANALYSIS LAKE PREP This is awesome for data scientists and businesses because... OUTPUT By combining traditionally separate tools – analytics, business intelligence, data visualization, data preparation and in-memory acceleration – a big data discovery platform reduces project completion from 12-18 months to a matter of hours. Slow and complicated THEN 0101010 0101010101 10 RAW)11 STRUCTURED DATA TRADITIONAL DATA OCDATAIO1 | 110 0010101 WAREHOUSE BI REPOSITORY 101010 10111 This used to hurt data scientists PREDETERMINED and businesses because... REPORTS ONLY Siloed data, expensive storage, myopic and irrelevant insights, and painfully slow last-generation tools hampered the ability of organizations to compete in a modern, on-demand marketplace. platfora 1 디디