Updated readme.

Author: Takanori Takase

README.md

@@ -8,4 +8,65 @@
 
 ## Overview
 
-(work in progress)
+**cryptowelder** is a cryptocurrency trade monitoring application to visualize the trading activities, positions, profits and losses in time series graph.
+
+1. Extract market data and account information from public APIs and private APIs.
+2. Transform the extracted data set into its generic data model.
+3. Store the transformed data into its time-series database.
+4. Visualize the time-series data with various graphs and charts. 
+
+![Grafana Dashboard Screenshot](./docs/img/dashboard.png)
+
+### Multi-Currency, Multi-Exchange
+Collect and store data from multiple exchanges for multiple products and multiple currencies. 
+Collected data sets are displayed in a single unified view.
+
+### Realtime Visualization & Notification
+Time-series data displayed graphs/charts, and are continuously updated in real-time as they are collected. 
+Displayed time-windows are configurable on the fly, such as today, last 24h, this week, last N days, and arbitrary time-points from T1 to T2.
+Configure custom alerts and notifications. (cf: "Email and/or Slack me if indicator X is Y")
+
+### Web UI
+Monitor trading activities from desktops and/or mobile with a standard web browser. 
+Drag & drop graphs and components to create custom dashboards.
+
+### Home Currency Conversion
+Convert multiple currencies in a single home-currency unit (cf: 1.2 CoinA + 2.3 CoinB + 3.4 CoinC -> 2345.6 JPY).
+Which currency to evaluate to, and how to apply the conversions are dynamically configurable. 
+
+### Trade Statistics
+Display trading activities such as daily/monthly trading p/l, N-days trading volume per accounts/products.
+Track trading activity statistics such as comparing last N-days trading volume as of today versus that of yesterday.
+
+
+## Mechanics
+The application consists of the following components:
+* Set of [Python](https://www.python.org/) scripts for scraping data via public/private APIs.
+* RDBMS ([PostgreSQL](https://www.postgresql.org/)) for storing the time-series data.
+* [Grafana](https://grafana.com/) for visualization and alerting of the time-series data.
+
+
+## Getting Started
+
+### Prerequisites
+* A REHL-equivalent linux machine with command line interface and direct internet access. Cloud-hosted VM is preferred with at least 2GB memory.
+* Python 3.4.x or later with pip. `pyenv` + `pyenv-virtualenv` are recommended.
+* PostgreSQL 10.x or later installation, or privileges to install if not already installed.  
+* Grafana 5.x or later installation. (Root privilege is not needed for Grafana installation.) 
+* Exchange access tokens for private API access. 
+* Basic knowledge of Linux for configuration and installation. 
+* Basic knowledge of Python and SQL for customization. 
+
+### Steps
+1. Install and configure PostgreSQL instance. 
+ 1. Create database `crytowelder`.
+ 2. Execute the DDL and DML scripts.
+ 3. Create database user `grafana` and grant read-only access.
+2. Configure and build the Python scripts.
+ 1. Install pip requirements.
+ 2. Prepare local configuration file with access tokens and scraping policies.
+ 3. Launch the script to start collecting and storing data into the database.
+3. Install and configure Grafana instance.
+ 1. Configure users and security policies.
+ 2. Import the dashboard template included.
+ 3. Access Grafana's web ui with a web browser.