Python3 based ASX data download and web application with basic features:

• ability to search by sector, keyword, movement, dividend yield or other attributes

• watchlist and sector stock lists, sortable by key metrics (eps, pe, daily change etc.)

• historical company metrics (cashflow, earnings, balance sheet) using YFinance

• bonds/commodities/cryptocurrencies and macro-economic datasets using investpy

• portfolio optimisation using PyPortfolioOpt

• Market and per-sector performance incl. top 20/bottom 20

• Virtual portfolios with trend visualisations, profit/loss by stock

• Stock point-scoring by rules over time, based on their price performance

• Visualisations provided by plotnine and matplotlib

This software is provided to you "as-is" and without warranty of any kind, express, implied or otherwise, including without limitation, any warranty of fitness for a particular purpose. In no event shall the author(s) be liable to you or anyone else for any direct, special, incidental, indirect or consequential damages of any kind, or any damages whatsoever, including without limitation, loss of profit, loss of use, savings or revenue, or the claims of third parties, whether or not the author(s) have been advised of the possibility of such loss, however caused and on any theory of liability, arising out of or in connection with the possession, use or performance of this software.

• Python 3.8+

• Django 3.2+ (recommended with latest Djongo)

• Djongo (python MongoDB for Django apps)

• Other requirements documented in requirements.txt

On the system to run the website, you'll need to install required software:

git clone https://github.com/ozacas/asxtrade.git
cd asxtrade
sudo pip3 install -r requirements.txt
cd src/viewer # for setting up the website

Next, you'll want to setup a database to store the ASX data: instructions for this can be found at the MongoDB website

Data since 2020 is loaded into an archive file, which it is recommended you load to help get you started:

  mongorestore --db=asxtrade --archive=data/asxtrade-20210618.bson

This will give you 2021 and 2020 ASX prices, World Bank and ABS free datasets. The scripts provided below can then be run to ingest historical prices. The above command may need to be altered with credentials to the database and authorisation to perform the restoration.

python3 manage.py migrate

python3 manage.py createsuperuser

export MPLBACKEND=Agg # recommended: all testing is done with this matplotlib backend plugin to integrate with Django server
python3 manage.py runserver # run on local dev. host

This application only works with daily data fetched after 4pm each trading day from the ASX website. It will take several hours per run. After the run, you must update the current month's pre-pivoted market data cache. The setup and daily run process is similar to:

# setup database password to use
export PASSWORD=<your db password here>

# run periodically (eg. quarterly) and at setup to update list of securities/companies/fundamentals
python3 src/asxtrade.py --want-companies
python3 src/asxtrade.py --want-isin
python3 src/asxtrade.py --want-details

# daily: fetch prices from ASX - configuration in config.json. Run after trading close each trading day
python3 src/asxtrade.py --want-prices

# daily: update cached data for the specified month (1..12) - needed for website operation
python3 persist_dataframes.py --month 3 --year 2021 --status INCOMPLETE|FINAL --dbpassword '$PASSWORD'

This new data may not appear in the application until the cache entries expire (or alternatively you can restart the server).

Feature	Thumbnail Picture
Portfolio watchlist
Stock view
Search by financial metric
Market sentiment
Performance by sector
Portfolio optimisation
World Bank API data: more than 10000 free datasets in over twenty subject areas

Tox is used to drive pytest to execute the unit test. Django manage.py test is not used. Github Actions is then used to invoke tox for CI/CD for the application and unit tests.

HTML unit test coverage reports can be produced using something like (from the root of the local repo):

$ export PYTHONPATH=`pwd`/src/viewer
$ pytest -c pytest.ini --cov-report html --cov-config=.coveragerc --cov=src .
$ ls htmlcov
# open htmlcov/index.html in your favourite browser

Use of the Django Debug Toolbar is strongly recommended, but it does have a known limitation which prevents usage during normal operation due to the use of pandas within asxtrade: see jazzband/django-debug-toolbar#1053 for details. For this reason, the MIDDLEWARE setting for the toolbar is commented out, except when needed.

Recent releases include support for API data integration using the excellent wbdata, persisting the data to parquet format in the mongo database for the app. A utility script is provided to download this data: but it works very slowly to be nice to the Internet (only 2 datasets per minute). At that rate it would take over a week to download the entire dataset of nearly 18000 dataframes! Only year-based datasets are currently supported.

python3 ingest_worldbank_datasets.py --help

This program builds an inverted index of the WorldBank data, to speed searching via Show -> World Bank data on the left-side. Three different analyses are possible:

1. A single metric for a single country, each year
2. A single metric for several countries
3. Many metrics for a single country

Each analysis comes with its own visualisations and features. More to come...

It is now possible, but entirely optional, to add company financial performance metrics (earnings, cashflow, balance sheet etc.) to an existing database:

python3 ingest_financials.py --help

Ingestion of financial metrics should be run when companies update the market on their performance (eg. mid-year and year-end). To be nice to Yahoo Finance API, it will take a couple of days to perform a complete run.

It has been observed that in some cases the data ingested can be wrong, particularly for ASX stocks. Use at own risk. This data is therefore considered experimental and pages using this data are labelled with a warning. If data is not ingested, then the 'Show financial performance' button on stock views (as well as data download) will 404. This data is not available for ETFs and some stocks at this time. It is planned to support search for companies by a metric.

You can also use this program to fill gaps - as yfinance supports up to five years of historical data: use the --fill-gaps option for this. It is recommended to always fetch current data using asxtrade.py but at least historical prices can be filled in with this approach. There are numerous limitations with this approach: many fields of a quotation are left unchanged as no data is available.

ABS provides several different APIs:

• Indicator API (headline financial/economic indicators eg. employment, CPI etc.)

• free API using pandasdmx: census data and other datasets accessible online also

Each of these is described below.

Dataframes can be ingested into MongoDB using:

  python3 ingest_abs.py --help

Some of the datasets can consume more than 8GB of memory during download so a 16GB machine is recommended for ingestion. Alternatively reduce the startPeriod to reduce the amount of historical data fetched.

ABS provides a free-registration associated with headline data. This includes CPI, Employment, Import/Export and Earnings data, amongst others. All that is required is to register for a free API key (which can take a week or two to arrive via email) and then choose ABS Headlines from the left. Note that this data is downloaded at the time of usage, so it does not ingest a-priori.

Registration for a key can be done at the ABS Website

Once you have this key, modify src/viewer/app/settings.py to include your new key at ABS_API_KEY setting.

Preliminary support through investpy is now available for visualising data associated with commodities, cryptocurrencies and world-wide bonds. Nothing needs to be configured for this, as data is downloaded if requested.

• I like to experiment: its how I learn. I built this using a weird combination of tech - RPI4, Mongo, Django, Python, Matplotlib just to see whether it could be done!

• Keep skills sharp and up-to-date

• Learn finance and the macroeconomic world

• Give back to the community and learn some wonderful tools

• Have fun!