{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cluster Discovery Services for Telecommunications"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Introduction/Business Problem"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Providing optimum network coverage is goal of each Network Planning Deportment of a telecommunication company. For this reason, companies tend to perform surveys, marketing, and outsourcing so that maximum profit be gained by minimum resource deployment. With the lake of such planning and management, the companies suffers from many problems including customer churn, poor allocation of resources, and insufficient of customer segmentation and so in creating packages.  \n",
    "Therefore, we propose *to establish a **framework of services that offers cluster mining** for existing and potential customers,  optimum network planning, package segmentation & development by leveraging their existing data along-with __geographical data__*.  \n",
    "We discussed the proposal with our existing client that was previously National Telecommunication Company and now gone private. During our various discussion the proposal has been discussed frequently with their project director and operational manager. They were interested and excited about it and were looking for a pilot project in short term of time by focusing just one issue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Requirements"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Telecommunication companies are well equipped with customer data and geographical area of their interests. Such data exist in CRM solutions that contains customer addresses as well as service usage behavior along-with their financial patterns. Moreover, Network planning and operation departments are also rich in capturing their deployed resources like base stations and cabinets. Therefore, leveraging the data of customers’ profile and behavior along-with deployed resources and financial outcome can provide a strong base for data scientists to answer some of their top most problems.  \n",
    "1. Identifying clusters of customer churns   \n",
    "2. Identifying the geographical areas that require resource optimization  \n",
    "3. Identifying clusters of problems that affects customer services \n",
    "\n",
    "We can leverage very useful data of existing ongoing CRM project as well as Network Operation Solution. CRM can provide us customer profile information including customer age, location, status and also usage information like subscription-type, average monthly bill, current subscription status, number of issues. From Network Operation, we can have resource data like BST, Cabinet capacity, location, and Network outages, and customer-issues-reference. Alongside, we will use spatial map to plot our data for clustering.  \n",
    "\n",
    "\n",
    "\n",
    "We plan to start by first establishing a clear gloal by the concent of project director i.e., \"why customer churn in Islamabad city during April to July 2018\". The goal is being established within small scope, data processing & availability, limited resources, and no-budget constraints.\n",
    "\n",
    "After setting the goal, we esbalished the following hypothesis:\n",
    "\n",
    "Customer churn occurs due to:\n",
    "\n",
    "> increase number of network issues  \n",
    "> bill payment  \n",
    "> nighbor influence  \n",
    "\n",
    "We collected customer and their subscription data from CRM and Faults and Resources data Operation Management System (OMS) for the given period for the specified city. CRM will provide customer profile information (customer-age, subscription-date, location, service-type) and usage data (monthly-bill, tickets, feedback). OMS will provide network resources information (BTS location, Capacity, number-of-outages).\n",
    "\n",
    "After collecting and refingment of the data, we will normalize the data to fit our model so that k-mean clustering alogrithm be applied by iteratively increameting number of clusters until adiquate numbers could be justfitied.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Foursquare services will be used to fetch the coordinates of network resources and customer address based location\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}