Hail Storms in the USA (1955-2019)

by Liam Wood

For Day 2 of Dashboard Week, we were presented with data from the National Oceanic and Atmospheric Administration (NOAA) looking at Hail Storms.

The task was to use the data (including spatial data) to create a dashboard using Alteryx’s reporting tools.

Data Prep

I decided to use the Initial Points dataset (ShapeFile) as it had dates from 1955 through to 2019 without any missing years..

The dataset contained a row for each hail storm, with fields for each date part, the state, values for the size of the hailstones and injuries and fatalities that occurred, as well as a spatial object for each storm. I combined the date and time fields in a Formula tool to produce a DateTime field.

I then used a Find and Replace to add Month names to the Month Numbers which were already present.

At this point, I added a Record ID which I could use to count the records when I aggregated to various levels.

Now I was ready to separate out my different areas of analysis.

I kept my analysis at a high level, and used a Summarize tool each time to aggregate to the appropriate level. On this occasion, I grouped by Year and the used Count Distinct of Record ID to find the number of hail storms for each year over the historical period.

Then came my first encounter with the Reporting Tools!

I used an Interactive Chart which I configured to create an Area chart. Once inside the configuration you must create a layer to start, followed by choosing a chart Type from the drop down menu, the orientation and then selecting what fields will be used as your X and Y axes. I also changed the colours within the Layer tab, added a title in the Chart tab and finally added an annotation using Notes. These reporting tools require a Browse tool to be able to view the chart, so it’s important to add that.

For the next chart I aggregated to each discrete month so I could look at the seasonality involved with hailstorms. I grouped by the Month Number and the Month Name and again counted the distinct record IDs. It was then important to Sort the Month Number Ascending so that the Months were in order before feeding the data into another Interactive Chart.

Although I pursued other analyses, the last chart that I created was looking at the top 5 states by number of hail storms in the latest year (2019). This time round I aggregated grouped by Year and State and then filtered the year to create a stream for 2019, and added a filter from the F output to create a 2018 stream. I sorted the 2019 stream descending and sampled to the first 5 records.

I then joined the 2 streams on State, so that I had the number of storms for each year side by side. I calculated the % difference between them using a Formula tool and then sorted ascending by the 2019 value. I noticed that the bar chart sorted in the opposite direction to the input so sorting ascending was important. Finally, I fed that into another Interactive Chart as a bar chart, with the % difference values as text on the bars.

The Report Layout

I followed the examples given within Alteryx to configure the layout, which involve the Join Multiple tool and the Layout tool. I decided to layout the 3 charts horizontally side-by-side, so I used a Join Multiple (as it accepts more than 2 inputs) with inputs from each of the Interactive Chart tools. This was then followed up by a Layout tool where you will see each of the charts in a dialog box and you’re able to check or uncheck them and reorder them. There is also an orientation option which allows you to choose whether they sit horizontally or vertically. You can also choose whether the charts spread

I just wanted to add a title and the data source and credits as a footer, either side of the charts. I used 2 separate Report Text tools to write the text for these and then used the Join Multiple and Layout tools again, but this time configuring the 3 elements vertically with the Header first and the Footer last.

Lastly, add a Render tool to output as a Temporary PDF and you have a report document ready to go!