crumby

A Flask based web analytics app


License
MIT
Install
pip install crumby==0.7.1

Documentation

Crumby

Crumby is an open source application for tracking and reporting visitor usage of a website. Crumby is a Flask application so it works with well known tools such as Apache, MySQL, and Python. Crumby includes a simple web-based frontend and a JSON compliant RESTful API for reporting. Queries are assigned to a public or private endpoint. Private queries are only accessible by authenticated users.

Architecture

architecture analytics

Crumby App

The Crumby app is developed with Python and Flask so it can be used with any server that supports the Web Server Gateway Interface (WSGI). The Crumby app includes a command line interface for administration, a reporting module, and a tracking module.

Crumby CLI

The command line interface is used to administer the Crumby app, try crumby -h from the terminal to print help on this command.

usage: crumby [-h] {adduser,deluser,env,geoip,init,run,users} ...

A command line interface for the Crumby web analytics app.

positional arguments:
  {adduser,deluser,env,geoip,init,run,users}
    adduser             add user - grant private query access
    deluser             delete user - revoke private query access
    env                 display Flask configuration values
    geoip               download latest geoip database
    init                create template config file
    run                 launch Crumby in development server
    users               view list of users

optional arguments:
  -h, --help            show this help message and exit

Reporting Module

Apart from querying the database directly, visit and event records can be viewed at a webpage or consumed through the JSON compliant RESTful API. The default locations for these resources are:

Webpage: https://<crumby.host>
API: https://<crumby.host>/api

Crumby ships with some built in queries which show common tracking metrics (e.g. visits per day, most popular pages, visitors location of origin). These queries are segmented into two groups: public and private. Public queries can be accessed by anyone. Private queries require authentication.

Tracking Module

The tracking module contains the cmb.js script and configures endpoints for receiving tracking data.

Crumby Database

A SQL database is required to store tracking data. Any database compatible with SQLAlchemy can be used. The following tables will be initialized when Crumby first runs:

  • calendar - A range of dates available for queries
  • events - Visitor interactions with the website
  • users - Users permitted to access private queries
  • visits - Visitors to the website

calendar

Column Category Description
datetime General Date and time

events

Column Category Description
id General Unique record ID
ip General IP address
cid General Visitor ID (cookie)
datetime General Date and time event occurred
doc_title Page Page title
doc_uri Page Page URI
name Action Name of the event
value Action Value of the event

users

Column Category Description
id General Unique record ID
username Login Username
password Login Hashed password

visits

Column Category Description
id General Unique record ID
ip General IP address
cid General Visitor ID (cookie)
datetime General Date and time page accessed
doc_title Page Page title
doc_uri Page Page URI
doc_enc Page Encoding of page
referrer General Referrer from client browser
_referrer General Referrer from request header
platform Hardware/Software Browser platform
browser Hardware/Software Browser name
version Hardware/Software Browser version
screen_res Hardware/Software Browser screen resolution
screen_depth Hardware/Software Browser screen depth
continent Geospatial Continent
country Geospatial Country
subdivision_1 Geospatial Subdivisions
subdivision_2 Geospatial Subdivisions
city Geospatial City
latitude Geospatial Latitude
longitude Geospatial Longitude
accuracy_radius Geospatial Geo accuracy
time_zone Geospatial Time zone
lang General Language from client browser
_lang General Language from request header

Geospatial Database (geoip.mmdb)

Crumby geolocates visits with the MaxMind geoip2 library and the GeoLite2 City database.

architecture diagram

Web Pages

Tracking visits to a webpage with Crumby is simple. Just add a pointer to the analytics server cmb.js script. The pointer should be included in the source code of any page that should be tracked, likely at the bottom of the body to allow other content to load first.

<script src="https://<crumby.host>/cmb.js"></script>

To track events (e.g. button click), make a call to cmb.event(<name>, <value>) when the event is triggered.

Getting Started

  1. Setup WSGI Hosting Environment

There are many ways to deploy flask applications like crumby. Refer to the Flask Deployment Options for several Hosted and Self-hosted options. At a minimum a WSGI server environment is needed to deploy the crumby app.

  1. Install a SQL Database
  • Install a database supported by SQL Alchemy
  • Install associated python database drivers (e.g. pymysql)
  1. Install Crumby

Installation will vary depending on the WSGI hosting environment selected.

For hosted options, you may need the package source code, which can be found on github:

    git clone https://github.com/bmweiner/crumby.git

For self-hosted options, you may want to install Crumby from PyPI:

    pip install crumby

Note: Crumby uses the flask-bcrypt package. This package requires Python Development Headers for installation, which may or may not be installed depending on the distribution. hint: sudo yum install python-devel

  1. Install the latest geoip database

     cd ~
     crumby geoip
    
  2. Initialize a config file

     crumby init
    
  3. Edit config file

     vi crumby.cfg
    
     # domain name of web server running crumby
     DOMAIN = 'myanalyticsserver.com'
     # uri of the database
     SQLALCHEMY_DATABASE_URI = 'mysql+pymysql://root@localhost/mydatabase'
     # absolute path to geoip database
     GEOIP2_DATABASE_NAME = '/Users/username/GeoLite2-City.mmdb'
     # proxy hops to remove, typically 0
     PROXY_COUNT = 0
     # domain name of the website crumby is tracking
     CROSSDOMAIN_ORIGIN = 'mywebsite.com'
     # complex random value used to sign things
     SECRET_KEY = 'mysecretkey'
     # use secure session cookie
     SESSION_COOKIE_SECURE = True
    

Current environment variables can be viewed with the crumby CLI:

    crumby env
  1. Set an environment variable named: CRUMBY_SETTINGS to the config file path

     export CRUMBY_SETTINGS=/Users/username/crumby.cfg
    
  2. Deploy Crumby

Follow instructions for the deployment option selected to deploy the crumby app.

Development Server

Crumby can be run in the Flask development server (not recommended for production) with the crumby CLI:

crumby run

Adding Queries

Crumby ships with a bunch of SQL queries which live in Crumby/templates/api under the private or public directory. Public queries can be viewed by anyone and private queries require authentication.

New queries can be added to the public and private directories. These queries will be automatically added to the crumby webpage and API. The following guidelines should be followed:

  • One text file per query with the extension .sql
  • The filename will be the name of the query
  • Queries must follow the syntax of the SQL database used

Refer to the Crumby Database section for table names, field names, and descriptions.

Advanced Queries

Queries are templates processed by Flask using the Jinja2 syntax. Refer to the Jinja2 Template Designer Documentation to customize your queries.

By default, every query is passed the following variables:

  • t0: Start Time, default today - 30 days
  • t1: End Time, default: today

These variables are created using query string parameters passed with a request to the crumby API. Refer to the crumby API https://<crumby.host>/api/public, for specifications on the API interface.

  • days, int. number of consecutive days to retrieve before today.
  • from, str in form YYY-MM-DD. start date of date range.
  • to, str in form YYY-MM-DD. end date of date range.

This is helpful if you want to constrain a query by time, for example returning the count of unique users within a certain date range:

SELECT date(datetime) as date, count(distinct cid) as users
FROM visits
GROUP BY date
WHERE date(datetime) between date("{{t0}}") and date("{{t1}}")

Managing Users

To view private queries, users must be authenticated. Users can be listed, added, or removed with the crumby CLI:

crumby users
crumby adduser name password
crumby deluser name