If you have followed the Installation Instructions and have the Daylio premium mode activated, you may continue with the usage instructions here.
This doc assumes you have activate the virtual environment and are in the app directory
$ cd PATH_TO_REPO
$ source venv/bin/activate
This section covers the main feature of this project. The later features require this Clean a CSV sections steps to be completed first.
Export the data on your mobile device.
You will now have a file named daylio_export.csv which contains all your entries from the most recent back to the first. See CSV format for more info.
Download the file on your computer and then move CSV to dayliopy’s configured CSV input location.
$ mv daylio_export.csv PATH_TO_REPO/dayliopy/var/data_in/daylio_export.csv
Use the commands below to read in the above export file, clean the data and write out a new CSV to the configured location.
Note that this will overwrite the existing output file. This should be fine though, since the input file always contains all data to date.
$ make csv
python -m dayliopy.clean_csv
Reading CSV: /Users/mcurrin/repos/daylio-csv-parser/dayliopy/var/data_in/daylio_export.csv
Replacing the activities column with multiple activity columns
Writing cleaned CSV to: /Users/mcurrin/repos/daylio-csv-parser/dayliopy/var/data_out/cleaned.csv
Example values for a row of the output CSV:
| timestamp | datetime | date | weekday_label | weekday_num | mood_label | mood_score | clean | cook | music | stressed | note |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1522444920.0 | 2018-03-30 23:22:00 | 2018-03-30 | Friday | 4 | average | 3 | 1 | 0 | 0 | 1 | Did a roadtrip. |
| 1522360440.0 | 2018-03-29 23:54:00 | 2018-03-29 | Thursday | 3 | happy | 4 | 0 | 1 | 0 | 0 |
Formatted using Markdown table tool.
You can now open the cleaned CSV in a CSV editor and start using it. If you want to get the data into a database or want to fit a stats model to your data, follow the instructions in the sections below.
$ make mood
python -m dayliopy.mood_report
mood_score
mean: 3.28
median: 3.00
average 926
happy 781
sad 269
amazing 84
horrible 49
Name: mood_label, dtype: int64
The fit_model.py script performs the following steps:
$ make fit
OLS Regression Results
==============================================================================
Dep. Variable: mood_score R-squared: 0.955
...
...
==============================================================================================
coef std err t P>|t| [0.025 0.975]
----------------------------------------------------------------------------------------------
cook 0.5807 0.104 5.581 0.000 0.376 0.785
clean 0.3392 0.131 2.597 0.010 0.083 0.596
music 0.2331 0.096 2.432 0.015 0.045 0.421
...
weekday_1 0.5355 0.111 4.840 0.000 0.318 0.753
weekday_2 0.7794 0.107 7.316 0.000 0.570 0.989
...
weekday_6 0.5772 0.125 4.622 0.000 0.332 0.822
month_2 0.6454 0.132 4.889 0.000 0.386 0.905
month_3 0.9950 0.123 8.068 0.000 0.753 1.237
...
month_12 1.5238 0.126 12.083 0.000 1.276 1.771
year_2017 0.6669 0.079 8.465 0.000 0.512 0.822
year_2018 1.4733 0.129 11.451 0.000 1.221 1.726
==============================================================================
Omnibus: 6.218 Durbin-Watson: 1.591
Prob(Omnibus): 0.045 Jarque-Bera (JB): 6.075
Skew: 0.206 Prob(JB): 0.0480
Kurtosis: 3.166 Cond. No. 20.0
==============================================================================
$ make fix
$ make typecheck
Make sure to follow the clean CSV section above otherwise this section will not work.
Use SQLite to create a database and a new table called daylio.
$ cd PATH_TO_REPO
Create new empty DB file if none exists. Use the project’s set up script to import from the default cleaned CSV file location.
$ make db
View the schema:
$ make schema
CREATE TABLE daylio(
"timestamp" TEXT,
"datetime" TEXT,
"date" TEXT,
"weekday_label" TEXT,
"weekday_num" TEXT,
"mood_label" TEXT,
"mood_score" TEXT,
"cook" TEXT,
"clean" TEXT,
"music" TEXT,
"stressed" TEXT,
"note" TEXT
);
SQLite’s default behavior it to set the affinity for each column to TEXT - see Datatypes in SQLite Version 3.
Numeric calculations may still be done as if the columns were numeric. You may change the column types if you wish, by altering the table, or by creating a table with the same name manually before doing the import.
$ make interactive
sqlite> -- The default most is csv, which is not pretty.
sqlite> .mode columns
sqlite> .headers on
sqlite> SELECT ROUND(AVG(mood_score), 2) AS avg_mood, SUM(cook) AS sum_cook FROM daylio;
avg_mood sum_cook
---------- ----------
3.27 58
sqlite> .quit
$ cd dayliopy
$ sqlite3 var/data_out/db.sqlite -header -column \
'SELECT mood_label, date FROM daylio LIMIT 5;'
mood_label date
---------- ----------
happy 2018-03-30
happy 2018-03-29
average 2018-03-28
happy 2018-03-27
average 2018-03-26
$ sqlite3 var/data_out/db.sqlite -header -csv \
'SELECT mood_label, date FROM daylio LIMIT 5;' \
> path/to/report.csv