DATA SCIENCE SERVICESBI & Analytics
Business Intelligence is a bullet train
It move fast, delivering sub-second query times and transporting data to end users in visually sleek formats. Organizations are beginning to adopt BI at every level, empowering both departmental teams and individuals to execute self-service queries in addition to larger enterprise projects. But just as the bullet train ﬁrst brings you to your stop, leaving you to travel further to your ﬁnal destination, so too does diagnostic BI fail to deliver you all the answers. Yes, you can analyze your data; yes, you can empower your business leaders. But in a rapidly changing marketplace, how do you know you’re making the right decisions for tomorrow’s customer?
Every minute there are 2 million search queries on Google, 277,000 tweets, and 100 million new emails sent. With the world’s data doubling every two years, Analytics and Big Data is transforming the way we do business. It helps us understand our world in a fundamentally new way that can transport us to new levels of knowledge with spectacular velocity. But asking questions of your data is only eﬀective if you know the right questions to ask, and that requires skilled data scientists to interpret the results of your critical business questions
Basket Market Analysis
Because of the massive data explosion in every industry, we turn to Machine Learning to help us ﬁll the gaps in what we can compute. With the Fractional Data Science oﬀering, SolidQ helps your team understand the standard process for Data Mining and the implementation of Machine Learning.
A typical Fractional Data Science engagement comprises the listed six points below. It starts with developing a business understanding and then asks what data corresponds to that understanding. SolidQ helps you prepare your data, model it, and then deploy using Machine Learning.
Our Data Science team excels at mentoring and training. In this ﬁrst step of Data Science adoption they walk your team through the tactics of data preparation and modelling. This is a classroom-style element designed for the entire team. Training topic can include Azure ML, R coding, SSAS, and Data Mining.
Data Preparation and Overview
After your team transforms the data and loads it into a SQL Server testing database, SolidQ works to discover more in-depth understanding of the data. By preparing several data sets that take into account various time frames (months, years) we can determine the distribution of values and prepare several additional computed variables. We then work to identify the relationship between the passage of time and the customers’ behavior.
Preparing and Evaluating Data Mining Models
We identify patterns within the data by using directed data mining techniques – Decision Trees, Naïve Bayes and Neural Networks. We prepare several models with diﬀerent data sets then prepare those models by applying diﬀerent algorithm parameters for each set. Because there is no way of knowing in advance how many models we will actually need, we limit this phase of development with time constraints.
Evaluating the Eﬃciency of the Models
It’s important to spend time determining if the deployed models need reﬁnement or become obsolete. We conduct all eﬃciency measures in a small, controlled data warehouse and analyze them with OLAP cubes or PowerPivot.
Deployment parameters are set by you according to your teams skills and project scope. SolidQ can prepare deployment reports and DMX queries in your OLTP application. We can also include churn results into your existing OLAP cubes.
Review of the data and models
SolidQ will review the data models with your team after deployment to encourage reﬁnement and learning.