datafluent 0.0.1

Build a better understanding of your data in PostgreSQL.

Data Fluent for PostgreSQL

Build a better understanding of your data in PostgreSQL.

The following shows an example report generated by this tool. It gives the numbers of rows, columns, bytes as well as human-friendly size counts for each table within a given PostgreSQL database.

The following shows the row count for every column that represents a date grouped by year and month.

Installation

On Ubuntu 20:

$ wget -qO- \
    https://www.postgresql.org/media/keys/ACCC4CF8.asc \
        | sudo apt-key add -
$ echo "deb http://apt.postgresql.org/pub/repos/apt/ xenial-pgdg main" \
    | sudo tee /etc/apt/sources.list.d/pgdg.list

$ sudo apt update
$ sudo apt install \
    git \
    python3-pip \
    python3-virtualenv \
    postgresql-13 \
    postgresql-client-13 \
    postgresql-contrib

On macOS:

$ brew install \
    git \
    postgresql \
    virtualenv

Then, regardless of platform, setup a virtual environment and install the following Python-based dependencies.

$ virtualenv ~/.fluency
$ source ~/.fluency/bin/activate
$ python3 -m pip install \
    csvkit \
    humanfriendly \
    ipython \
    openpyxl \
    Pandas \
    psycopg2-binary \
    typer \
    xlsxwriter

Then clone this repo as it contains the datafluent_pg.py script:

$ git clone https://github.com/marklit/datafluent_pg.git ~/datafluent_pg
$ cd ~/datafluent_pg

Example Analysis

Clone fivethirtyeight's data repo. It has a large number of CSV-formatted datasets.

$ git clone https://github.com/fivethirtyeight/data.git ~/538data

Make sure you can access a PostgreSQL database on your machine. Here I'm creating an intel database for the mark user on my Ubuntu 20 machine.

$ sudo -u postgres \
    bash -c "psql -c \"CREATE USER mark
                       WITH PASSWORD 'test'
                       SUPERUSER;\""

With PostgreSQL access setup, create a database called intel.

$ createdb intel

I'll import one of the datasets within fivethirtyeight's repo. Note, because the dates within this dataset are not formatted in YYYY-MM-DD format, I needed to override the format so that the MM/DD/YYYY format would be read properly.

$ csvsql --db postgresql:///intel \
         --insert ~/538data/congress-generic-ballot/generic_topline_historical.csv \
         --datetime-format="%m/%d/%Y"

I'll run the Excel Report Generator:

$ python datafluent_pg.py

This will result in a fluency.xlsx file being produced with two worksheets: Metrics and Time Distributions.

If you need to override any parameters, please refer to the documentation:

$ python datafluent_pg.py --help

Usage: datafluent_pg.py [OPTIONS]

Options:
  --pg-dns TEXT                   [default: postgresql://localhost:5432/intel]
  --output TEXT                   [default: fluency.xlsx]
  --install-completion [bash|zsh|fish|powershell|pwsh]
                                  Install completion for the specified shell.
  --show-completion [bash|zsh|fish|powershell|pwsh]
                                  Show completion for the specified shell, to
                                  copy it or customize the installation.

  --help                          Show this message and exit.