Data analysis of SMHI data
In the project we have produced script using Bash, C++ and ROOT to clean up raw data, analyze it and make some plots. The report can be found as the pdf report.pdf. The project is divided in three parts, as described in the report.
Organisation of Directories
In directory CleanData you can find the directory datasets, in which all the raw unprocessed datasets from SMHI are found. The code directory, the code for cleaning up the data necessary for Project A and Project B are found. The code for doing analysis and making plots is there as well. Everything should run automaticaly if you follow the procedure described below. When you start root, processed data will be filled put in the ProcessedData/UppsalaData dir, and plots will be put it in the images directory.
Getting Started
To be sure that everything works, the script has been tested on the following versions:
Use root version 5.34/30
On Lubuntu (32-bit)
C++ version: 5.4.0 2016/06/09
Note that if the rootlogon file runs (that is, if you start run in the code directory), a different version of root might produce strange results, such as weird plots and wrong output. Make sure to use the right version of root.
Prerequisites
What things you need to install the software and how to install them
Bash, Ubuntu, C++, root in the above mentioned versions.
Order of running scripts
To run Project A and C, simply cd into code directory and type
.L root -l
The plots should be shown and saved in /images and /plots.
To run Project A only:
If you want to extrapolate to a certain year, call the function tempYearplotandpred(int yearToExtrapolate) while in root (it should be compiled in rootlogon.C but if not, just type .L tempYearplotandpred.cpp+ while in root in code directory).:
tempYearplotandpred(int yearToExtrapolate)
This will output the predicted annual average temperature of that year, based on a polynomial fitted to the whole data set. The reduced chi-square of the fit is also output to standard output, it should be around 0.8, not so bad in other words.
To run Project B only:
To run Project B that shows the number of days that have an daily average temperature below 0 in each year, run the file in the directory code/ named cleaning_data.sh (WARNING! The code is slow and it can take up to 30 min to finish!!)
To run Project C only:
To run Project C that plots the distribution of the hottest and coldest date in a year, type the following lines after you are in ROOT under directory code/:
.L tempTrender.cpp+ // if it is not loaded
tempTrender t("../CleanData/datasets/uppsala_tm_1722-2013.dat");
t.hotCold_Upp();Then the plots would be save in the directory images/.
Authors
- Leo Zethraeus
- Piotr Yartsev
- Xi-Zhen Liu