Key Requirement: Be able to dynamically verify and compare different site weather activity.
Delivered: Data cleansing and dashboard using Power BI.
Below is an example of an interactive Power BI report of over 25 million underlying hourly sample records, pivoted and recalculated on the fly according to your selections. Data is according to terms from the UK Environmental Change Network ECN run by the Centre for Ecology and Hydrology and is set to disallow download of all base data so it remains protected. These regular meteorological samplings are from their (and partner) AWS – Automated Weather Stations. Please visit their website, which also has notes regarding any intermittent quality issues with the recordings and data.
You can expand the Power BI charts using the button on bottom right. It looks like a simple set of static charts, but select the left hand bars to silhouette / filter the main charts which then show comparisons against overall averages. Most other items are also interactive, which is where true BI reporting gets it’s power.
Note the Slicers (to sub-select country, AWS, Year-Month etc) have been moved to new collapsible Filters in the recent November 2018 release – they work excellently in both Desktop and PowerBI.com but have to vote up my suggestion to get it to work in this “publish to web” facility!
Exploratory Data Analysis (EDA)
Power BI with data cleansing
Z-Score consolidation of non-homogenous metrics for comparison of variances
Visualisation shows patterns enabling conclusions on dependencies