COMM 205 MIDTERM REVIEW SESSION BY KEVIN DHIR

Transcription

1 COMM 205 MIDTERM REVIEW SESSION BY KEVIN DHIR

2 Page 1 of 19 TABLE OF CONTENT 1. TABLE OF CONTENTS 2. INTRODUCTION 3. IF & NESTED IFS 4. AND/OR 5. COUNTIFS, SUMIFS 6. VLOOKUP 7. LEFT, RIGHT, MID, & CONCAT 8. LEN, TRIM, REPLACE, SUBSTITUTE 9. FIND, SEARCH 10. STATA 11. TAB, SUM, GEN, BYSORT, GSORT 12. TAB, SUM, GEN, BYSORT, GSORT (continued) 13. MISSING VALUES 14. MERGING DATASETS 15. MERGING DATASETS (continued) 16. TOSTRING, DESTRING, BROWSEIF 17. EXCEL PRACTICE 18. STATA PRACTICE

3 Page 2 of 19 INTRODUCTION Kevin Dhir, Second Year in Combined Major Business & Computer Science o Stata is cool but has applications that most of you won t use in the future o Excel is dope and at the end of this course combined with what you learned last year in COMM290, you ll feel like a WIZARD o Did pretty well in COMM205 (86%) last semester This package is an overview of the content in the course, condensed into a twohour review session. You will not succeed by just attending this review session, but you will hopefully have a better understanding of the many functions! This review session is INTERACTIVE. Play around with Excel and Stata as we go! One of the keys to success for this exam is the cheat sheet. There are many small syntax tricks possible and the cheat sheet will save you from those. I ll be going over how to use formulas and commands together, interchangeably, the logic behind some things, how to approach certain questions, and providing you a few sample problems that I have generated. I ll have office hours on INSERT DATE AND TIME. Timeline 1. EXCEL Functions minute break 3. Stata Commands 4. Cheat Sheet Tips 5. Q&A, Sample Questions You will need to know the following Excel functions: IF, AND, OR, COUNTIFS, SUMIFS, VLOOKUP, CONCATENATE, LEN, TRIM, SUBSTITUTE, REPLACE, FIND, SEARCH You will need to know the following Stata commands: tab, sum, rename, browse, gen, drop[if], keep[if], bysort, duplicates tag, merge, egen, destring, tostring, browse if

4 Page 3 of 19 IF & NESTED IFS CHEATSHEET MATERIAL: Syntax: =IF(logical_test, [value_if_true], [value_if_false]) Logical Tests: = (equal to), <> (not equal to), < (less than), <= (less than or equal to) > (greater than), >= (greater than or equal to) IF is used to test for specific conditions and return either true or false The basic concept is quite clear, but when the real trickiness comes when they are NESTED. It helps to draw out a tree diagram to follow a nested IF function. For instance, assuming A1 contains GR: =IF(A1>55, CR, IF(A1<50, F, D )) SERIOUSLY, DRAW THE TREE. Strings within an IF are surrounded by quotation marks. Do not use quotation marks around cell references, formulas, or numerical values.

5 Page 4 of 19 AND, OR CHEAT SHEET MATERIAL: =AND(logical1,[logical2], ) =OR(logical1,[logical2], ) AND and OR are logical operators. For AND: For OR: If all logical tests are true, then AND will return TRUE. If ANY of the logical tests are true, then OR will return TRUE. THESE CAN BE USED IN COMBINATION WITH IF! So you can do IF-AND and IF- OR. REMEMBER THIS! Again, TREES ARE YOUR FRIENDS. Now you try! Recreate the below spreadsheet in Excel. I typed a formula using AND in C2, OR in D2, and a nested AND in E2. I then dragged it down. What three formulas did I use? Nesting these within an IF allows you to print certain strings or values based on the results, as opposed to true.

6 Page 5 of 19 COUNTIFS, SUMIFS CHEAT SHEET MATERIAL: =COUNTIFS(criteria_range1, criteria1, ) (up to 127) =SUMIFS(sum_range, criteria_range1, criteria1 ) (up to 127) In this course we only deal with COUNTIFS and SUMIFS. These are NOT the same as COUNTIF and SUMIF. Those only allow for a SINGLE criterion. These allow for between 1 and 127 criteria. You may have a question around this! Think of the criteria as an AND(). If, and only if, all the criteria are true, it will either count or add the value to the sum. The sum_range MUST be a numerical range. You can have your sum_range and your criteria_range be the same range. The two must have the same size and shape. COUNTIFS counts the number of times the criteria given evaluate to true. Criteria format does not NEED logical operator if you are just checking if it is equal to. But they can also be preceded by an equal sign (if they do, then quotation marks are required. Criteria constraints must be surrounded by, even if it is a logical operation on a numerical. THIS IS DIFFERENT THAN WITH REGULAR IF! For COUNTIFS, can have MULTIPLE Criteria Ranges but must be same size! EXAMPLE! D12 =SUMIFS(A11:A14,B11:B14,"BUCS") C14 =COUNTIFS(A11:A14,">90")

7 Page 6 of 19 VLOOKUP Open your eyes, VLOOKUP to the skies and see Freddie Mercury CHEAT SHEET MATERIAL: =VLOOKUP(lookup_value,table_array,col_index-num, [range_lookup]) VLOOKUP looks in the first column of the table_array for the lookup_value, then returns the value found within the column index number. Write a VLOOKUP function to get Stanley s Bank Account Balance. Write a VLOOKUP function to get Evan s Mark on Test. Write a VLOOKUP function to get the name of the student with balance of 888. Why should you NOT use column B as your lookup table? EXTRA: Write the string SNAKES in cell A8 using LEFT/RIGHT/CONCAT/MID TRAPS WITH VLOOKUP Looks up value in the FIRST column then uses col index RELATIVE to that. If using an approximate match, the first column of the table_array must be sorted in ascending order Each value in the first column must be unique, else it will just return the first one it comes across If looking up through a range, the range values must only be the lower bound of the range (e.g. Range: 90, Grade: A+ instead of Range , Grade A+

8 Page 7 of 19 LEFT, RIGHT, MID, CONCATENATE CHEAT SHEET MATERIAL: =LEFT(text,num_chars) =RIGHT(text,num_chars) =MID(text,start_num,num_chars) =CONCATENATE(text1,[text2],...) =text1&[text2] These three functions are used to manipulate text, and return substrings (text within text). NOTE: THE START_NUM OF MID IS INCLUSIVE. Cell A1 contains the string YR2MDHIR. The first three characters specify the student s year standing, the fourth specifies their gender, and the final characters are their last name. What will the following functions return? =LEFT(A1,3) =RIGHT(A1,5) =LEFT(RIGHT(A1,4),2) =MID(A1,3,1) =MID(A1,4,1) =LEFT(A1,500) =RIGHT(A1,500) =MID(A1,14,1) Concatenate allows you to join strings together. It can be used interchangeably with &. We want to rewrite the text to instead say DHIR YR2 M. Use LEFT, MID, and RIGHT to create substrings, then them together. If you wish to add spaces, simply and. This does not have to be referenced from a cell.

9 Page 8 of 19 LEN, TRIM, SUBSTITUTE, REPLACE CHEAT SHEET MATERIAL: =LEN(TEXT) =TRIM(TEXT) =SUBSTITUTE(text,old_text,new_text,[instance_num]) =REPLACE(old_text,start_num,num_chars,new_text) LEN returns the number of characters in a string, inclusive of blank characters (spaces). TRIM removes repeated blanks (spaces). Great for data cleaning to remove errors. WHY WOULD THIS HAPPEN!? THEY BOTH SAY KEVIN! SUBSTITUTE vs REPLACE. SUBSTITUTE is used when you know the specific string you are trying to swap out! What if we wanted to change M to say MALE? WHAT WENT WRONG IN B2?! We have to specify the INSTANCE! What would be written in C1? REPLACE is great for inserting strings into other strings, or replacing the text at a certain point in a string. Unlike SUBSTITUTE, it doesn t know exactly what it will replace, just where. Two big tricks! With SUBSTITUTE, you must know the instance number you are trying to replace. With REPLACE, you must pay careful attention to the start position and number of characters

10 Page 9 of 19 FIND, SEARCH CHEAT SHEET MATERIAL: =FIND(find_text,within_text,[start_num]) =SEARCH(find_text,within_text,[start_num]) SEARCH and FIND sound and look super similar, but have a few KEY differences. SIMILARITIES: Return position of find_text relative to start_num If start_num > within_text s length, returns #VALUE! If start_num is negative, returns #VALUE! If find_text is not found, returns #VALUE! Start_num is assumed to be 1 SEARCH Not CaSe SENSitiVE ALLOWS Wildcard Characters FIND CASE SENSITIVE NO WILDCARD CHARACTERS WILDCARDS? -> can be used to replace a single character, e.g. p?t will match with any three letter string beginning with a p and ending in a t (pet, pat, pot, pit, etc.) * -> can be used to replace any (including 0) number of characters, e.g p*t will match with poot, part, port, parrot, etc ~ -> if you are looking for a * or? in your find_text, you must precede it with a ~ to stop it from being treated as a wild card (e.g. to find the string who?, you would have to denote it who~? Otherwise it would match with whom, etc) USES FIND and SEARCH are usually used in combination with MID and REPLACE to give a starting point to replace a certain string or crop a certain string

11 Page 10 of 19 STATA THE STATA SECTION OF THE EXAM DEPENDS SO HEAVILY ON YOUR CHEAT SHEET. MAKE YOUR CHEAT SHEET YOUR BEST FRIEND. BASICS SYNTAX IS KEY I cannot stress enough. Your cheat sheet is going to be your BIBLE when it comes to syntax. EVEN if you think you know the syntax, I would strongly suggest you double check any commands you re asked to evaluate or write. That s a great place or way for them to pull a fast one. GENERAL SYNTAX: command varname(s) [if varname==value] [, options] EVERYTHING IN STATA IS CASE SENSITIVE. THIS INCLUDES COMMANDS, VARNAMES, ETC. The Log File Allows you to save the output of Stata without permanently editing your dataset. Must be initiated EVERY time you use Stata if you care about your results. The Do File Basically a macro. Allows you to complete multiple operations by opening just one file. Can save time and be used to avoid typos. Datasets Upon opening a dataset in Stata, the variables window will show you all the different columns in the dataset. The properties window will give you details about specific variables that you select, such as the # of observations. Let s explore a dataset

12 Page 11 of 19 TAB, SUM, GEN, EGEN, BYSORT, GSORT tabulate, tab, or ta produces a frequency distribution table for the variable. Mainly used for CATEGORICAL variables (discrete). Option: sort, defaults to descending, summ Frequency, Percent, Cumulative Frequency summarize, sum, or su produces number of observations, average(mean), standard deviation, max, and min value of a NUMERICAL variable. SYNTAX TRAP: Make sure you enclose any text string that s not a COMMAND or VARIABLE NAME in quotation marks, such as specific variable values ex. sum ch if loc== USA

13 Page 12 of 19 Combining tab and sum Allows user to see a frequency distribution table for each variable within a specific varname o E.g. tab loc, sum(emp) -> generates the frequency distribution table of employees for each location o Does NOT allow the use of the detail option generate, gen, or g creates a new variable in the dataset. Mainly used for simple math assignments using operators such as +, -, *, / Can gen things like the sum of two other variables, a constant, a string THERE ARE STRICT RULES ON THE VARIABLE NAMES POSSIBLE: CASE SENSITIVE, 32-CHARACTER LIMIT, FIRST CHARACTER CANNOT BE A NUMBER<CANNOT CONTAIN SPACES OR OTHER SPECIAL CHARACTERS SUCH AS * & % etc o If you want to use an already taken name, consider using the rename function. It has the SAME restrictions on naming. egen is a much more powerful generate that uses built in functions. Trust in the cheat sheet Trust in your mind. Most common functions: mean, median, max, min, count, total Generally used in combination with bysort, o For the purposes of COMM 205: Vertical? Use egen. Horizontal? Use gen. bysort is named in reverse. Think of it as sort-by! Sort your dataset by the given variable(s). Useful for breaking down a large dataset. For COMM205, only use egen&sum gsort sounds super cool like bysort, but in reality it s just the ability to sort in either ascending or descending order. sort by multiple variables effect if there exist duplicate values for the first var.. gsort -primarykey +randomvar will be the same as gsort primarykey-.

14 Page 13 of 19 MISSING VALUES Two types: Missing NUMBERS -> denoted by a period (.) in the cell o Though it may not be relevant to the exam, numbers are in black text while strings are in red text o Texts in blue are numerical variables that are disguised as strings (recall the _merge variable) o Numerical MVs are treated as positive infinity, so keep this mind when used with if greater than statements! Missing STRINGS -> denoted by a blank cell Count number of missing observations using summ if command o For numerical -> summ fyear if emp==. will show the number of missing observations in the dataset o For string-> summ fyear if tic== CLEAR, REPLACE, DROP, KEEP These four commands are used almost exclusively in combination with if. Replace o Changes the variable s value, usually based upon a condition such as if the value is not == a specific number, or if it is a missing value Clear o Removes the active / master dataset that is currently loaded into memory If you don t clear before opening another, it will simply open both together Drop/drop if o Deletes an entire column from your dataset (drop) or database, or drops certain rows based on the condition given (drop if) Keep/keep if o The complement of drop, often used to retain only non-missing values e.g. keep if emp!=. will then drop every row that is missing emp

15 Page 14 of 19 MERGING DATASETS duplicates checks if a value given in the first varlist listed contains any duplicates, then generates a new variable in which a numerical 0 or 1 is printed based upon if there was a duplicate or not o to check the number of duplicates tab the var you generated (usually tab) There are THREE types of merging. Familiarize yourself with the concept of a PRIMARY KEY -> unique variable or set of variables that can be used to identify an observation in a data set. One-to-One o Combines the datasets using one or more key variables to link the observations between the data sets o Both datasets have the same primary key o The variable names and variable types must be identical! One-to-Many o The primary key of the master dataset appears on multiple lines of the using dataset o E.G. master dataset = class list, contains var stu_num, using dataset = grades list, contains vars grade ass_num and stu_num. Many instances of stu_num in the using

16 Page 15 of 19 Many-to-One o The master dataset contains multiple observations containing the primary key of the using dataset Direct opposite of one-to -many SIX STEPS FOR MERGING TWO DATASETS Step 1: Determine the primary key (unique identifier) of the master dataset o This can be one or multiple variables Step 2: Determine the primary key of the using dataset o This is the dataset that is not currently loaded into your Stata Step 3: Determine which variables that you will be using to merge the two datasets o They must exist in both datasets in the exact same format (name and type), this INCLUDES CASING! Step 4: Is the primary key of the master dataset also the variable(s) you will be using to merge? Yes or No o If the primary key consists of multiple variables, then this is only yes if you are using ALL of said variables to merge Step 5: Is the primary key of the using dataset also the variable(s) you will be using to merge? Yes or No Step 6: Choose from the below o If Step 4 is YES and Step 5 is YES, it is 1:1 (Primary Key of Master & Using) o If Step 4 is YES and Step 5 is NO, it is 1:m (Primary Key of Master) o If Step 4 is NO and Step 5 is YES, it is m:1 (Primary Key of Using)

17 Page 16 of 19 TOSTRING, DESTRING, BROWSE IF tostring converts a numerical variable into a string variable o Can either convert the variable in place (using the replace option) or generate a new variable (using the generate option) o NOTE: numerical missing values, represented by a period, will be turned into the string. Instead of a blank! You can fix this using drop if. destring converts a string variable into a numerical variable (what a shock) o again, can either replace or generate a new variable o Doesn t have the same problem that tostring does with missing values o If the string isn t a number it ll simply put in a missing value browse if allows you to view specific observations based on a condition o Useful for analyzing data manually, identifying outliers etc. o Remember: missing numericals are considered to be positive infinity THAT S A WRAP! HERE ARE SOME PRACTICE PROBLEMS.

18 Page 17 of 19 EXCEL PRACTICE PROBLEMS Bob is an aspiring Sauder School of Business Student. He one day dreams of being a billionaire. He is attending a networking event, which has released the following information on its delegates. The actual spreadsheet has 500 values. Bob writes a command to calculate the number of delegates attending who have a net-worth greater than $2 million who are in the Tech field. What formula does he use? =COUNTIFS(B11:B15,">2000",C11:C15,"Tech") Bob really respects these fine gentlemen. He wishes to edit Column A to prefix each name with the string Mr.. What formula does he use? =REPLACE(A11,1,0,"Mr. ") Bob connected with a fellow named Suge Nathair. He wants to find out what field he is working in. He searches for his name in Column A, and wants the cell he writes the formula in to print out the field he is working in. What formula does he use? =VLOOKUP( Suge Nathair, A11:A500, 3, 0)

19 Page 18 of 19 STATA PRACTICE PROBLEMS Generate a new variable which will display the shareholder s equity of each firm. What formula would you use? gen sh_equity = asset/liability firm asset liability A 5 2 B 4 3 C 8 3 What command would you use to clean outliers from this dataset? You try to use tab and sum together, and write the command tab airport_name, sum (numberofpassengersinmillions) if year == Stata keeps giving you an error. How do you fix this? You want to sort the dataset in descending order of number of passengers and from oldest year to newest year. How would you do this?