ACCT 420: Course Logistics
Session 1
Dr. Richard M. Crowley
About Me
Teaching
- Third year at SMU
- Previously taught ACCT 101
- Before SMU: Taught at the University of Illinois Urbana-Champaign while completing my PhD
Research
- Accounting disclosure: What companies say, and why it matters
About this course
What will this course cover?
- Foundations
- Learning the ropes of R
- In class: Getting down the most important skills
- Outside: Practice and refining skills on Datacamp
- ~4 hours in week 1 and 2
- Financial forcasting
- Predict financial outcomes
- Linear models
Learning and getting familiar with R and forecasting
What will this course cover?
- Binary classification
- Event prediction
- Classification/detection
- Advanced methods
- Non-numeric data
- Anomaly detection
- AI/Machine learning
- 2 weeks on current developments
Using R for higher level financial forecasting and detection
Datacamp
- Datacamp is providing free access to their full library of analytics and coding online tutorials
- You will have free access for 6 months (Usually $29 USD/mo)
- Online tutorials include short exercises and videos to help you learn R
- I have assigned materials via a Datacamp class, which will count towards participation
- Check your email or eLearn for access
- Datacamp automatically records when you finish these
- I have personally done every assigned tutorial to verify their quality
- You are encouraged to go beyond the assigned materials – these will help you learn more about R and how to use it
Datacamp’s tutorials teach R from the ground up, and are mandatory unless you can already code in R.
Textbook
- There is no required textbook
- Datacamp is taking the place of the textbook
- If you prefer having a textbook…
- R for Everyone by Jared Lander is a good one
- Other course materials (slides and articles) are available at:
- eLearn
- https://rmc.link/acct420
- Announcements will be only on Elearn
Teaching philosphy
- Analytics is best learned by doing it
- Less lecture, more thinking
- Working with others greatly extends learning
- If you are ahead:
- The best sign that you’ve mastered a topic is if you can explain it to others
- If you are lost:
- Gives you a chance to get help the help you need
- If you are ahead:
Grading
- Standard SMU grading policy
- Participation @ 10%
- Individual work @ 30%
- Group project @ 30%
- Final exam @ 30%
Participation
- Come to class
- If you have a conflict, email me
- Excused classes do not impact your particpation grade
- If you have a conflict, email me
- Ask questions to extend or clarify
- Answer questions and explain answers
- Give it your best shot!
- Help those in your group to understand concepts
- Present your work to the class
- Do the online exercises on Datacamp
Outside of class
- Verify your understanding of the material
- Apply to other real world data
- Techniques and code will be useful after graduation
- Answers are expected to be your own work, unless otherwise stated
- No sharing answers (unless otherwise stated)
- Submit on eLearn
- I will provide snippets of code to help you with trickier parts
Group project
To be announced later
Final exam
- Why?
- Ex post indicator of attainment
- How?
- Likely only 2 hours
- Long format: problem solving oriented
- Potentially a small amount of MCQ
- When?
- Tentatively set for Tuesday, Dec 4 @ 1pm
Expectations
In class:
- Participate
- Ask questions
- Clarify
- Add to the discussion
- Answer questions
- Work with classmates
- Ask questions
Out of class
- Check eLearn for course announcements
- Do the assigned tutorials on Datacamp
- This will make the course much easier!
- Do individual work on your own (unless otherwise stated)
- Submit on eLearn
- Office hours are there to help!
- Short questions can be emailed instead
Tech use
- Laptops and other tech are OK!
- Use them for learning, not messaging
- Examples of good tech use:
- Taking notes
- Viewing slides
- Working out problems
- Group work
- Avoid:
- Messaging your friends on Telegram
- Working on homework for the class in a few hours
- Watching livestreams of pandas or Hearthstone
Office hours
- Walk-in hours from 10:30-11:30am Fridays
- Or by appointment
- Short questions can be emailed
- I try to respond within 24 hours
About you
About you
- Survey at rmc.link/aboutyou
- Results are anonymous
- We will go over the survey next week at the start of class
Introduction to analytics
Learning objectives
- Theory:
- What is analytics?
- Application:
- Who uses analytics? (and why?)
- Methodology:
- Introduction to R
*Almost every class will touch on each of these three aspects
What is analytics?
What is analytics?
Oxford: The systematic computational analysis of data or statistics
Webster: The method of logical analysis
Gartner: catch-all term for a variety of different business intelligence […] and application-related initiatives
What is analytics?
Simply put: Answering questions using data
- Additional layers we can add to the definition:
- Answering questions using a lot of data
- Answering questions using data and statistics
- Answering questions using data and computers
Made using seancarmody/ngramr
Analytics vs AI/machine learning
- In class reading:
- AI Will Enhance Us, Not Replace Us
- By DataRobot’s Senior Director of Product Marketing
- Shortlink: rmc.link/420class1
How will Analytics/AI/ML change society and the accounting profession?
What are forecasting analytics?
- Forecasting is about making an educated guess of events to come in the future
- Who will win the next soccer game?
- What stock will have the best (risk-adjusted) performance?
- What will Singtel’s earnings be next quarter?
- Leverage past information
- Implicitly assumes that the past and the future predictably related
Past and future examples
- Past company earnings predicts future company earnings
- Some earnings are stable over time (Ohlsson model)
- Correlation: 0.7400142
Past and future examples
- Job reports predicts GDP growth in Singapore
- Economic relationship
- More unemployment in a year is related to lower GDP growth
- Correlation of -0.1047259
Past and future examples
- Ice cream revenue predicts pool drownings in the US
- ???
- Correlation is… only 0.0502886
- What about units sold?
- Correlation is negative!!!
- -0.720783
- What about price?
- Correlation is 0.7872958
This is where the “educated” comes in
Forecasting analytics in this class
- Revenue/sales
- Shipping delays
- Bankruptcy
- Machine learning applications
What are forensic analytics?
- Forensic analytics focus on detection
- Detecting crime such as bribery
- Detecting fraud within companies
- Looking at a lot of dog pictures to identify features unique to each breed
Forensic analytics in this class
- Fraud detection
- Working with textual data
- Detecting changes
- Machine learning applications
Forecasting vs forensic analytics
- Forecasing analytics requires a time dimension
- Predicting future events
- Forensic analytics is about understaninding or detect something
- Doesn’t need a time dimension, but it can help
These are not mutually exclusive. Forensic analytics can be used for forecasting!
Who uses analytics?
In general
- Companies
- Finance
- Manufacturing
- Transportation
- Computing
- …
- Governments
- AI.Singapore
- Big data office
- “Smart” initiatives
- Academics
- Individuals!
53% of companies where using big data in a 2017 survey!
What do companies use analytics for?
- Customer service
- Royal Bank of Scotland
- Understanding customer complaints
- Royal Bank of Scotland
- Improving products
- Siemens’ Internet of Trains
- Improving train reliability
- Siemens’ Internet of Trains
- Their business
- $18.3B USD market in 2017
- Just a small portion of overall IT spending ($3.7T USD)
- $18.3B USD market in 2017
What do governments use analytics for?
- Govtech
- Open data
- AI Singapore
- Talent matching
- AI in health Grand Challenge
- AI research funding
What do academics use analytics for?
- Tweeting frequency by S&P 1500 companies (paper)
- Aggregates every tweet from 2012 to 2016
- Shows frequency in 5 minute chunks
- Note the spikes every hour!
- The white part is the time the NYSE is open
What do academics use analytics for?
- Annual report content that predicts fraud (paper)
- For instance, discussing income is useful
- first row is decreases, second is increases
- But if it’s good or bad depends on the year
- For instance, in 1999 it is a red flag
- And one that Enron is flagged for
What do individuals use analytics for?
- Consulting
- Radim Řehůřek: Maintainer of gensim, freelance consultant
- Investing
- Health
- Smart watches and other wearables
Why should you learn analytics?
- Important skill for understanding the world
- Gives you an edge over many others
- Particularly useful for your career
- Jobs for “Management analysts” are expected to expand by 14% from 2016 to 2026
- Accountants and auditors: 10%
- Financial analysts: 11%
- Average industry: 7%
- All figures from US Bureau of Labor Statistics
Introduction to R
What is R?
- R is a “statistical programming language”
- Focussed on data handling, calculation, data analysis, and visualization
- We will use R for all work in this course
Why do we need R?
- Analytics deals with more data than we can process by hand
- We need to ask a computer to do the work!
- R is one of the de facto standards for analytics work
- Third most popular language for data analytics and machine learning (source)
- Fastest growing of all mainstream languages
- Free and open source, so you can use it anywhere
- It can do most any analytics
- Not a general programming language
Programming in R provides a way of talking with the computer to make it do what you want it to do
Setup
- For this class, I will assume you are using RStudio with the default R installation
- RStudio downloads
- R for Windows
- R for (Max) OS X (Download R-3.5.1.pkg)
- R for Linux
- You will need a laptop or desktop for this
- I am working to find a lab on campus for this as well
- For the most part, everything will work the same across all computer types
- Everything in these slides was tested on R 3.5.0 and 3.5.1
How to use R Studio
- R markdown file
- You can write out reports with embedded analytics
- Console
- Useful for testing code and exploring your data
- Enter your code one line at a time
- R Markdown console
- Shows if there are any errors when preparing your report
How to use R Studio
- Environment
- Shows all the values you have stored
- Help
- Can search documentation for instructions on how to use a function
- Viewer
- Shows any output you have at the moment.
- Files
- Shows files on your computer
Basic R commands
Arithmetic
- Anything in boxes like those on the right in my slides are R code
- The slides themselves are made in R, so you could copy and paste any code in the slides right into R to use it yourself
- Grey boxes: Code
- Lines starting with
#are comments- They only explain what the code does
- Lines starting with
- Blue boxes: Output
# Addition uses '+'
1 + 1## [1] 2# Subtraction uses '-'
2 - 1## [1] 1# Multiplication uses '*'
3 * 3## [1] 9# Division uses '/'
4 / 2## [1] 2Arithmetic
- Exponentiation
- Write xy as
x ^ y
- Write xy as
- Modulus
- The remainder after division
- Ex.: 46 mod 6=4
- 6×7=42
- 46−42=4
- 4<6, so 4 is the remainder
- Integer division (not used often)
- Like division, but it drops any decimal
# Exponentiation uses '^'
5 ^ 5## [1] 3125# Modulus (aka the remainder) uses '%%'
46 %% 6## [1] 4# Integer division uses '%/%'
46 %/% 6## [1] 7Variable assignment
- Variable assignment lets you give something a name
- This lets you easily reuse it
- In R, we can name almost anything that we create
- Values
- Data
- Functions
- etc…
- We will name things using the
<-command
# Store 2 in 'x'
x <- 2
# Check the value of x
x## [1] 2# Store arithmetic in y
y <- x * 2
# Check the value of y
y## [1] 4Variable assignment
- Note that values are calculated at the time of assignment
- We previously set
y <- 2 * x - If we change the values of
xandyremain unchanged!
# Previous value of x and y
x## [1] 2y## [1] 4# Change x, then recheck the value
# of x and y
x <- 200
x## [1] 200y## [1] 4Application: Singtel’s earnings growth
Set a variable
growthto the amount of Singtel’s earnings growth percent in 2018
# Data from Singtel's earnings reports, in Millions of SGD
singtel_2017 <- 3831.0
singtel_2018 <- 5430.3
# Compute growth
growth <- singtel_2018 / singtel_2017 - 1
# Check the value of growth
growth## [1] 0.4174628Recap
- So far, we are using R as a glorified calculator
- The key to using R is that we can scale this up with little effort
- Calculating every public companies’ earnings growth isn’t much harder than calculating Singtel’s!
Scaling this up will give use a lot more value
- We can also leverage functions to automate more complex operations
- There are many functions built in, and many more freely available
- We’ll cover this next week
- We’ll also need ways to read data files and work with collections of numbers
- We’ll cover this next week as well
Wrap up
- R Practice
- Shortlink: rmc.link/420r1
- Do the practice here if you would like help with it
- Otherwise, do it at home
- For next week:
- Start working on the Datacamp tutorials!
- Assigned tutorials are on the Datacamp class page
- For next week, complete the Intro to R course
- More tutorials will be assigned in future weeks
- Other helpful tutorials:
- Start working on the Datacamp tutorials!
1
ACCT 420: Course Logistics
Session 1
Dr. Richard M. Crowley
