One of the biggest misconceptions about analytics is that it’s a software package you have to purchase. In fact, analytics is the process of using data to solve business problems – and is really about the marriage of an important business question, the right data set and an analyst who knows the techniques needed to answer the question at hand.
In many cases, the systems we use to track our business already include reports that answer our business questions, but if not, analysts rely on a few basic tools to help managers make better decisions with data.
So what are these tools?
1. A Tool for Extracting Data
Many of the applications I work with, such as Salesforce.com, include some mechanism for extracting data for external use. Usually, this is done through some sort of “export to Excel” function.
In other cases, I might choose to pull data directly from the databases that hold application data, and may choose to use complicated data integration tools such as Informatica or SQL Server Integration Services to do so.
2. A Relational Database for Storing and Cleaning Your Data
If that data comes from more than one system, or there’s a lot of it, or the work I am doing is going to be repeated over and over again, I generally opt for a relational database to store and manage my data sets.
The upshot of doing so is that I can use languages such as SQL to fix problems in my data sets, such as common misspellings or incorrect accounting system rollups. Depending on the complexity of the problem I might build a single-use database with only a couple of tables, or I might choose to design and deploy a complicated data warehouse.
The hard truth is that most relational database engines generally do the same things, and the right choice is usually whichever one you already own. I tend to prefer Microsoft’s SQL Server because it’s easy to use, reasonably priced and ships with several add-ons that create a lot of value for me, but Microsoft Access, Oracle, IBM’s DB2 and MySQL are also strong choices.
3. A Reporting Tool to Create Attractive Charts and Graphs
There are dozens of easy to use tools in the marketplace for creating attractive reports, charts and graphs from the data you’ve chosen for analysis – and hundreds that aren’t so great. Again, like the relational database question, the right tool for the job is probably the one you already own and like using.
To get my arms around the problem I’m trying to solve, I often start with Microsoft Excel – it’s easy to use, powerful and nearly all of my clients own it, and with the addition of PowerPivot, can create almost any type of chart and graph I like to use. Lately, though, I’ve favored next generation rapid visualization tools such as Tableau, QlikView and Spotfire. In general, these tools will connect directly to relational databases and can help an analyst figure out what kind of picture best tells his or her story.
In many cases, this is where I stop. Sometimes, though, I need to create a report that can easily be refreshed by end users and distributed throughout an organization. In cases like that, I usually take what I learn from the rapid analysis tool and implement it in whichever enterprise reporting platform my clients prefer.
Microsoft clients might be on Sharepoint’s Power View or SQL Server Reporting Services; others might have chosen SAP BusinessObjects or IBM’s Cognos. Again, products in this class all tend to do similar things and the right tool for the job is the one that your clients will use.
4. A Data Mining or Statistics Package for the Harder Stuff
Then, there’s the hard stuff: the data mining and predictive modeling that is typically outside of the tasks most organizations would choose to tackle themselves.
Projects such as fraud analysis, business experimentation and forecasting often start with the same kinds of data sets used for basic reporting analysis, then run them through statistical algorithms to find new and exciting insights. In general, these kinds of projects are more dependent on selecting the right technique than the software used to create the analysis, though finding the right tool can greatly streamline the whole process.
To tackle such problems, analysts typically turn to data mining and statistics packages such as IBM’s SPSS, a host of products from SAS or the open source statistics language R. A word of warning, though: Proper use of these techniques and tools typically requires the kind of specialized training pursued by Ph.D. candidates, scientific researchers, mathematicians and data scientists.
DataClear is a Baton Rouge-based data analytics consulting firm. Contact Us for a free 30-minute consultation and discover how your company can profit from data-driven decision making using tools that won’t break your budget.