Tough time we are facing due to the pandemic. Now is the time to invest in our infrastructure, whether to meet increase demand patterns or decrease. Data cultures are essentials to maintain efficient operations in both cases.
Whether you are pro at keeping data ready for analysis using data warehouses or still using excels. Achieving a data competent culture is still a challenge. We are in the age of self-serve business intelligence where we expect all our staff to be involved and revolving around data.
What are the first steps of making use of data?
We need to create general statistics about this data. Data here must naturally be organized, aka transformed, to be fit for analysis. First, we will gather the data, Extract, then we will apply the transformation to it, and then we will load it to make use of it. This is known as ETL process: Extract, Transform, and Load. The purpose of this stage is to get our data in the right format organized into a tabular format, i.e. tables with columns with a unified data type. Numbers with numbers, text with text, and date with date. "apples with apples, oranges with oranges".
2. Descriptive Analytics
Now we can apply typical analysis to our data. Let us find sums of sales, average price, maximum and minimum fluctuations, finding deviation from standard parameters, etc.
We are aggregating our data. Classical tools are pivot tables and pivot charts.
In this stage, we are trying to describe our data. We are making sense of data. Typically this data is in the past.
This level gives us clarity on what happens in the fast. Which products have the highest sales? When was our peak patient admittance? How much did we produce? etc.
Recent years have witnessed an increase is data visualization tools, freely available, such as Microsoft Power BI, and Google Data Studio, and many more.
3. Diagnostic Analytics
This is the natural progress of what we can do with our data. We can dig deep into our data by adding drilling parameters. This is typically done by adding multiple layers to our data. In data warehouses this is can be easily done as the data is already structured to be analyzed. In cases where we are using datasheets, for example: Excel, we will need to sort the data. Perhaps unpivot tables, classify data, apply fuzzy merge, etc.
One we get the data into multiple dimensions we can simply do this step. It will give us causal relationships of the parameters we got from previous steps. For example which production line gave us the slowest production? Which product had the most delays? When did we sell the most out of product X? What were the common symptom of the patients admitted?
This phase asks the question, why did this event happen?
For more details please join our dedicated online course. Contact us for more details.
For feedback and inquiry about the article contact the author.