What are all stages in datascience methodology?

Data science the rapidly evolving stream that employs different strategies to come up with an answer for existing business problem is a methodical approach that involves the following phases to start with understanding the business issue until upto coming up with an answer for business problem.
Here are the many different stages of data science methodology:
1) Understand business issues
2) Determining analytical approach
3) Determining data requirements for building analytical models
3) Collect data from many different datasources
4) Understand the type of data. This can be from relational databases, website logs, structured as well as unstructured data
5) Prepare data
6) Modelling data and continuous evaluation of models
7) Deployment of models designed to predict outcomes
8) Get feedback from customers and implement changes to models as needed
What does a typical job of a data analyst or someone who works closely with customers to understand business requirements and utilize data science to solve their business issue do?
A data analyst professional helps client to understand and improve the user experience with their online properties. For collecting data some datasource is needed. This can typically be firm’s website as well as web based or desktop based applications associated with storing of data like a EMR system in a typical hospital environment
Clients look for automated analytical solutions that will help them look at the metrics in form of dashboard views, reports etc. A data analyst will work with developer and automate this process using reporting solutions like SSRS, tableau etc depending on the solution being used for product
Data analyst is sometimes expected to do SAS programming that automates the insight retrieval based on statistical modelling techniques like significance testing, t-test, regression analysis etc. These techniques help with pre-treatment, post-treatment, control analysis to identify the best performing location in case of website projects. This can be header, top banner parallel to header, side banner to name a few. The location prominence helps clients place important information based on user behavior for maximum conversion which will be the primary business issue. This is typically referred to as user experience
Propensity matching Technique is the Technique used to predict a set of customers likely to have similar characteristics as another customer group. This technique is typically used in case of projects demanding market segmentation as opposed to projects that involve yes or no type questions
Model to Predict the customers who are more likely to sign up and activate for a product that will be launched in future. This is future business prediction that helps clients with decision making on production as well as inventory

Bigdata Hadoop Hortonworks Articles FREE:

Delivered by FeedBurner