Dashboard Week: Monday

by Valerija Kirjackaja

For DS21’s Dashboard Week, Andy chose the theme of Environment.  For our first project on Monday, we were asked to look at water temperature changes in the Gulf of Mexico. The data set contained observations gathered from several agencies around the Gulf of Mexico and covered the time period between 1922 and 2014.

After downloading the data and exploring it in Tableau, I decided to focus on changes in average and maximum observed temperatures. I noticed, however, that some of the MAX values were extremely high, so I decided to have a look at the dataset in Alteryx as the file contained over 8 million rows. After sorting temperature values in a descending order, I saw that some records showed 3 and 4-digit temperatures, which was clearly too high for sea water, so I decided to exclude those. The challenge was to establish the correct cut-off point for exclusion as there were also temperatures showing 99, 78, 50 degrees etc. The measurement column said all values were supposed to be in Celsius, so after doing some additional research on highest water temperature in the region, I decided to only keep values under 40 degrees. I also excluded negative temperature values and observations between 1922-1973, as well as year 2014, as there were too few records in those years.

After bringing the new hyper file into Tableau, I quickly rejected the idea of trying to build any maps, and exploring which locations have seen the highest rise/drop in temperatures over time as data processing was taking too long, even working with an extract. So, I decided to focus on a couple of charts that would show the spread of observations best: heatmap, barbell chart and a box and whisker plot.

For the barbell chart and the box and whisker plot, I looked specifically at the difference between 1973 and 2013, and found that although the average yearly temperature only increased by 1 degree Celsius, the difference in highest observed temperatures was more significant.

I also included two line charts, one showing the overall trend in average, highest and lowest temperature over the 40 years, and the other one looking at monthly patterns across all years.

Avatar

Valerija Kirjackaja