程序代写代做 ER database 7CCSMDDW: Coursework 2

7CCSMDDW: Coursework 2
Lecturer: Natalia Criado
Coursework assigned: 11 November 2019
Coursework deadline: 4:00 pm, 09 December 2019
Late Submissions deadline capped at 50: 4:00 pm, 10 December 2010
Overview
The purpose of this coursework is to create a normalised database and a data warehouse given a dataset. The coursework is formally assessed and is worth 10 of your final grade and is broken down to 10 points distributed below. You will receive some feedback as part of the marking of the coursework.
Part 1: Data Loading, Transformation Normalization 7 out of 10 points
1.1 Backup Restoring 0.5 points
1.2 Data Loading 1.5 points
1.3 Data Preprocessing 3 points
1.4 Data Integration 0.5 points
1.5 Database Normalization 1.5 points
Part 2: Constraint Implementation 2 out of 10 points 2.1 Status Consistency 1 point
2.2 Status Transition 1 point
Part 3: Data Warehouse 1 out of 10 points
3.1 ER Diagram 0.5 point 3.2 SQL Statements 0.5 point
Requirements
Working with raw crime data from a police department involves analysing a sequence of records about crimes. There are two different tables in this dataset. A table with crime data and a table with images of crime locations. The data is shipped in three different files: i crimes2015.sql containing information about crimes reported in 2015 and the images of the locations of all crimes, ii crimes2013.txt containing information about crimes reported in 2013, and iii crimes2014.csv containing information about crimes reported in 2014.
Each record in the crime table represents an crime, with the following information:
date reported: date when the crime was reported to the police date occ: date when the crime occurred
time occ: time when the crime occurred
1

occurred: the timestamp when the crime occurred. This column replaces the date occ and time occ in the crimes2015.sql file.
dr no: crime id
area: code of the area where the crime occurred
area name: name of the area where the crime occurred
rd: road number
crime no: code of the type of crime
crime desc: description of the type of crime
status: code of the crime status
status desc: description of the crime status
image no: code of the image of the place where the crime occurred
Each record in the image table contains the image of a crime location:
image no: image id
image: image stored a the BLOB value it is a binary large object that can hold a variable
amount of data
In a big city, there is a large number of crimes, so being able to read and analyse the data efficiently is very important. For this assignment you will import raw crime data into a MySQL database. As is typical of big data, there are many inconsistencies and anomalies in the dataset, so you may need to preprocess and clean the data before it is usable.
Part 1: Data Loading, Transformation Normalisation
Setup. In Part 1, you will use SQL commands to load, transform and normalise the crime data. On KEATs you will find a .zip file which contains template SQL files to edit for your Part 1 submission.
Be sure to:
1. Include your NAME at the top of the file in a SQL line comment.
2. Edit these files as text, not Word files or propriety SQL software.
3. Do NOT rename the files.
4. Include any comments using the SQL line comment i.e. lines beginning with .
Submission and Evaluation. Include your edited versions of these files in the .zip of your Part 1 submission. We will evaluate these SQL files using NMSs server which is version 5.5.50 MariaDB precisely. As a student, you can utilize NMSs server recall it is accessible via: https: nms.kcl.ac.ukmysql or your own database server installed on your personal computer to test your files and commands.
You can assume that these files are executed in the order that they appear in the description that follows:
1. backupRestoring.sql 2. dataLoading.sql
3. dataPreprocesing.sql
2

4. dataIntegration.sql 5. normalization.sql
Any file that is missing, renamed, or does not run will result in 0 points for that section.
Part Part
1.1: Backup Restoring. Write the required SQL statements to restore the data con tained in the file crimes2015.sql in the provided template file: backupRestoring.sql
1.2: Data Loading. Write the SQL statements to perform the following operations:
1.2.1. Table Creation. Write the required SQL DDL Data Definition Language state ments i.e. CREATE TABLE to create two tables named crimes2013 and crimes2014 that can held the data contained in the datasets crimes2013.txt and crimes2014.csv. En sure that:
table and attribute names do not conflict with SQL reserved words
attribute data types are core SQL data types as described in the lectures
1.2.2. Data Load. Write the required SQL statements to load the data contained in the datasets crimes2013.txt and crimes2014.csv into the tables crimes2013 and crimes2014.
Write all of these SQL statements in the above order in the provided template file: dataLoading.sql
1.3: Data Preprocessing. Write the required SQL statements to remove any duplicates, inconsistencies and anomalies in the data that is in the tables crimes2013, crimes2014 and crimes2015. Among others, the following inconsistencies and anomalies need to be solved:
The tables in the different datasets do not have the same number of columns.
The column time occ in some of the tables may not be formatted properly e.g., the
value 1 should be interpreted as 00:01.
The column image no in some of the tables may contain invalid values.
The capitalization of the status codes status are not consistent in the three tables.
Different formats may be used for the dates.
The data about some area names are missing.
Write all of the SQL statements required to preprocess the data in the provided template file:
dataPreprocesing.sql
1.4: Data Integration. Write the required SQL statements to integrate the data in the tables crimes2015, crimes2014 andcrimes2013 in a single table named crimes. Write these SQL statements in the provided template file: dataIntegration.sql
1.5: Database Normalization. Write the required SQL statements to normalize the crimes table into a set of tables in 3NF. Ensure that:
every table has a primary key
all foreign keys are properly declared
Write these SQL statements in the provided template file: normalization.sql 3
Part
Part
Part

Part 2: Constraint Implementation
Setup. In Part 2, you will use SQL commands to implement domain constraints in the tables contained in the crime2015.sql backup. On KEATs you will find a .zip file which contains template SQL files to edit for your Part 2 submission. Again, be sure to:
1. Include your NAME at the top of the file in a SQL line comment.
2. Edit these files as text, not Word files or propriety SQL software.
3. Do NOT rename the files.
4. Include any comments using the SQL line comment i.e. lines beginning with .
Submission and Evaluation. Include your edited versions of these files in the .zip of your Part 2 submission. We will evaluate these SQL files using NMSs server which is version 5.5.50 MariaDB precisely. As a student, you can utilize NMSs server recall it is accessible via: https: nms.kcl.ac.ukmysql or your own database server installed on your personal computer to test your files and commands.
You can assume that these files are executed in the order that they appear in the description that follows:
1. backupRestoring.sql 2. constraints.sql
Any file that is missing, renamed, or does not run will result in 0 points for that section. Write the SQL statements to perform the following operations:
Part 2.1: Status Description Consistency. Create MySQL stored programs e.g., trigger to ensure that the status and status desc are consistent e.g., it is not possible to have two different status descriptions for the same status code.
Part 2.2: Status Transition. Create MySQL stored programs e.g., trigger to ensure that the status code of crimes behaves as indicated by the following transition diagram:
CREATE
UNK IC
AO AA
JO JA
That is the crimes are created with the status UNK, then crimes can transition to status IC, from that status they can transition to AO or JO, etc. Crimes can only be deleted when they are in status AA or JA.
Write all of these SQL statements in the above order in the provided template file: constraints.sql 4

Part 3: Data Warehouse
Setup. In Part 3, you will create a data warehouse to analyse the number of crimes contained in the crimes2015.sql backup. On KEATs you will find a .zip file which contains template SQL files to edit for your Part 3 submission. Again, be sure to:
1. Include your NAME at the top of the file in a SQL line comment.
2. Edit these files as text, not Word files or propriety SQL software.
3. Do NOT rename the files.
4. Include any comments using the SQL line comment i.e. lines beginning with .
Submission and Evaluation. Include your edited versions of these files in the .zip of your Part 3 submission.
1. ER Diagram. Submit your ER Diagram as a PDF document not a Word document for compatibility reasons.
Use a diagramming program such as Microsoft Visio, OneNote, PowerPoint, Illustrator etc… for you ER diagram. Do not submit handdrawn diagrams, as they are usually unreadable! You can also use web apps such as http:draw.io to help you sketch out your diagrams.
Include your pdf in the .zip of your Part 2 Coursework submission.
2. SQL Statements. In Part 3, you will also develop SQL statements to create a data warehouse.
Include your edited version of the file warehouse.sql in the .zip of your Part 3 submission.
We will evaluate these SQL files using NMSs server which is version 5.5.50MariaDB pre cisely. As a student, you can utilize NMSs server recall it is accessible via: https: nms.kcl.ac.ukmysql or your own database server installed on your personal computer to test your files and commands.
You can assume that these files are executed in the order that they appear in the description that follows:
a backupRestoring.sql b warehouse.sql
Any file that is missing, renamed, or does not run will result in 0 points for that section.
Part 3.1: EntityRelationship Diagram. The police department would like to perform dif ferent queries related to the occurrence of individual crimes, by type, area, status, and by when they occurred and when they where reported. Given the data available, design a data warehouse that meets these requirements. Draw the ER Diagram of the star schema of this data warehouse. Your diagram must also cover the requirements below:
Include all entities, attributes including data types, and relationships. If necessary, please explainjustify your design choices.
5

Cardinalities and Assumptions. On the ER Diagram include all cardinalities of the relationships using the ER Diagram notation from the lecture. State assumptions made and ensure that those assumptions do not contradict with the coursework requirements. Please list your assumptions as bullet points.
Constraints. i For each relation, identify their primary and foreign keys. Simply add PK and FK to indicate the attributes that serve as primary and foreign keys, respectively.
Part 3.2: SQL Statements. Write the SQL statements to build and populate the data ware house in accordance with your ER Diagram. Note that the police would like to keep the data they have already stored in the crimes2015.sql file; you should ensure that the police does not lose any data as a result of the creation of the data warehouse.
Write your SQL statements in the provided template file: warehouse.sql
6