Hi everyone!
This is my first blog post, so let me introduce myself. I’m Charles Verdon (@chverdon), Technology Solutions Professional for Data Platform and Business Intelligence at Microsoft for Eastern Canada. I’m passionate about data analytics and creating insights through visualizations.
With all the news articles and general fear about Ebola, I wanted to better understand the spread of the epidemics with a Power BI dashboard. Within a few hours, here’s the results I obtained and the key insights that I popped out:
The mortality rate of confirmed and probable cases is actually trending down from a peak of 64% to about 46% probably from the help of concerted international response teams and investments.
Liberia has seen a big increase in new cases and deaths in late July surpassing Sierra Leone.
Total cases and deaths seem to be accelerating, confirming media reports that Ebola is moving faster than current containment efforts. This is further illustrated by the recent cases in Texas, Spain, and Saudi Arabia.
Data Source
To create this dashboard, I browsed the World Health Organization’s website and found that they release Disease Outbreak News press releases every few days that contain data tables such as this one:
Confirmed, probable, and suspect cases and deaths from Ebola virus disease in Guinea, Liberia, Nigeria, and Sierra Leone, as of 16 August 2014
New (1) | Confirmed | Probable | Suspect | Totals (by Country) | ||
Guinea |
| |||||
Cases | 24 | 396 | 140 | 7 | 543 |
|
Deaths | 14 | 252 | 140 | 2 | 394 |
|
Liberia |
| |||||
Cases | 48 | 200 | 444 | 190 | 834 |
|
Deaths | 53 | 178 | 202 | 86 | 466 |
|
Nigeria |
| |||||
Cases | 3 | 12 | 0 | 3 | 15 |
|
Deaths | 0 | 4 | 0 | 0 | 4 |
|
Sierra Leone |
| |||||
Cases | 38 | 775 | 34 | 39 | 848 |
|
Deaths | 17 | 326 | 34 | 5 | 365 |
|
Totals |
| |||||
Cases | 113 | 1383 | 618 | 239 | 2240 |
|
Deaths | 84 | 760 | 376 | 93 | 1229 |
|
(1) New cases were reported between 14 and 16 August 2014. |
| |||||
Data Transformation
The first thing I had to do was to manipulate this input data to split the columns and keep only the records needed for analysis. To automate this, I used Power Query and then consolidated the result to a table:
Partial data excerpt from two press releases consolidated
The second part was to use Power Query again to unpivot this data in a format ready for analysis:
The last step in creating the analytics model is to add the formula for the mortality calculation in Power Pivot:
Mortality:=sum([Deaths])/sum([Cases])
Visualizations
Designing visualizations in Power View is now as simple as creating Pivot Tables. You can go to the Insert tab and select Power View. Afterwards, you can select the fields required in the chart with the pane to the right of Excel.
Deploying to Power BI
The ultimate step is to save this spreadsheet to Office 365 Power BI using the Save to SharePoint Online feature. Once the spreadsheet is saved, users can now connect to the Power BI Portal and look at the interactive dashboards through any web browser:
Once enabled, users can also use Power BI Q&A to type questions and receive interactive answers in the form of charts and graphics:
Total ebola death by country and classification
And that’s it. With these simple steps anyone can use Microsoft Power BI to gain valuable insights!
Try a Power BI for O365 trial subscription on http://www.PowerBI.com.
Learn more about Power BI: www.aka.ms/SQLevents
Try Power BI for O365 trial subscription here www.PowerBI.com
Take Power BI lab http://blogs.technet.com/b/cansql/archive/2014/08/18/18-free-sql-server-2014-azure-and-bi-virtual-labs.aspx