There are often times when programmers need to produce test data for a system under development. They may generate some strings of random characters. Or they may pull in some text from a lorem ipsum generator. Both of these options are undesirable in that they do not look real. Non-developer testers of the system may look at a user interface populated with random characters or gibberish and feel that something is broken. This post explores using random number generators and data look-up tables to produce realistic-looking test data.

Background

When playing a tabletop, pen-and-paper role-playing game, such as Dungeons & Dragons, sometimes a game master needs help coming up with ideas for the world they are creating and describing to the other players. A common practice is to use tables of data covering various topics that can be randomly selected from using a set of dice. For example, if you need to figure out what is hidden inside a treasure chest, you could consult a treasures table. Or if you need to come up with the name for a character, you could consult a character name table. There are all sorts of tables covering a tremendous number of topic areas that can be used in this way. There are even role-playing games that create the majority of their play scenarios using this type of approach. Now, on to the dice used to make these tables work.

The types of dice used for role-playing games are special. Yes, you can use traditional six-sided dice that you may be used to from a game such as Monopoly or Yahtzee. But there are many other shapes and types of dice, known as polyhedral dice, that are useful in role-playing games. Here are some common polyhedral dice types used:

Name Description Link to Roll
d4 A four-sided die which gives results from 1 to 4. Roll!
d6 A traditional six-sided die which gives results from 1 to 6. Roll!
d8 An eight-sided die which gives results from 1 to 8. Roll!
d10 A ten-sided die which gives results from 1 to 10. The numbering can also be from 0 to 9. In that case, the 0 is treated as the 10. Roll!
d12 A twelve-sided die which gives results from 1 to 12. Roll!
d20 A twenty-sided die which gives results from 1 to 20. Roll!
d100 A 100-sided die (or two d10s can also be used) which gives results from 1 to 100. Roll!

(Table #1: Types of Dice)

Equipped with this information, let’s look at some real-world scenarios.

Generating Names

Let’s say you need to create a list of real-looking names. Using this handy list of first names from the Social Security Administration, you could create a table of 100 first names (first fifty are male, second fifty are female) like this:

First
Name
—roll d100
1. Michael 2. Christopher 3. Matthew 4. Joshua 5. Jacob
6. Nicholas 7. Andrew 8. Daniel 9. Tyler 10. Joseph
11. Brandon 12. David 13. James 14. Ryan 15. John
16. Zachary 17. Justin 18. William 19. Anthony 20. Robert
21. Jonathan 22. Austin 23. Alexander 24. Kyle 25. Kevin
26. Thomas 27. Cody 28. Jordan 29. Eric 30. Benjamin
31. Aaron 32. Christian 33. Samuel 34. Dylan 35. Steven
36. Brian 37. Jose 38. Timothy 39. Nathan 40. Adam
41. Richard 42. Patrick 43. Charles 44. Sean 45. Jason
46. Cameron 47. Jeremy 48. Mark 49. Stephen 50. Jesse
51. Jessica 52. Ashley 53. Emily 54. Sarah 55. Samantha
56. Amanda 57. Brittany 58. Elizabeth 59. Taylor 60. Megan
61. Hannah 62. Kayla 63. Lauren 64. Stephanie 65. Rachel
66. Jennifer 67. Nicole 68. Alexis 69. Victoria 70. Amber
71. Alyssa 72. Courtney 73. Rebecca 74. Danielle 75. Jasmine
76. Brianna 77. Katherine 78. Alexandra 79. Madison 80. Morgan
81. Melissa 82. Michelle 83. Kelsey 84. Chelsea 85. Anna
86. Kimberly 87. Tiffany 88. Olivia 89. Mary 90. Christina
91. Allison 92. Abigail 93. Sara 94. Shelby 95. Heather
96. Haley 97. Maria 98. Kaitlyn 99. Laura 100. Erin

(Table #2: First Names)

And then, using this list of most common last names from ThoughtCo, you could create a table of 100 last names like this:

Last
Name
—roll d100
1. Smith 2. Johnson 3. Williams 4. Brown 5. Jones
6. Garcia 7. Miller 8. Davis 9. Rodriguez 10. Martinez
11. Hernandez 12. Lopez 13. Gonzales 14. Wilson 15. Anderson
16. Thomas 17. Taylor 18. Moore 19. Jackson 20. Martin
21. Lee 22. Perez 23. Thompson 24. White 25. Harris
26. Sanchez 27. Clark 28. Ramirez 29. Lewis 30. Robinson
31. Walker 32. Young 33. Allen 34. King 35. Wright
36. Scott 37. Torres 38. Nguyen 39. Hill 40. Flores
41. Green 42. Adams 43. Nelson 44. Baker 45. Hall
46. Rivera 47. Campbell 48. Mitchell 49. Carter 50. Roberts
51. Gomez 52. Phillips 53. Evans 54. Turner 55. Diaz
56. Parker 57. Cruz 58. Edwards 59. Collins 60. Reyes
61. Stewart 62. Morris 63. Morales 64. Murphy 65. Cook
66. Rogers 67. Gutierrez 68. Ortiz 69. Morgan 70. Cooper
71. Peterson 72. Bailey 73. Reed 74. Kelly 75. Howard
76. Ramos 77. Kim 78. Cox 79. Ward 80. Richardson
81. Watson 82. Brooks 83. Chavez 84. Wood 85. James
86. Bennet 87. Gray 88. Mendoza 89. Ruiz 90. Hughes
91. Price 92. Alvarez 93. Castillo 94. Sanders 95. Patel
96. Myers 97. Long 98. Ross 99. Foster 100. Jimenez

(Table #3: Last Names)

To use these tables, you would role a d100 three times. The first number gives you a first name from table #2, the second number gives you a middle initial from table #2 (just use the first letter of the name), and the third number gives you a last name from table #3. Let’s try that. (If you don’t have a d100, use the Roll link above in Table #1.)

  • Roll #1 (d100) for first name (Table #2): 57
  • Roll #2 (d100) for middle initial (Table #2): 2
  • Roll #3 (d100) for last name (Table #3): 84

Ok, let’s map those numbers to the tables. “57” in table #2 gives: “Brittany”. “2” in table #2 gives: “C.”. And “84” in table #3 gives: “Wood”. For a full name of: “Brittany C. Wood”. Let’s try another.

  • Roll #1 (d100) for first name (Table #2): 29
  • Roll #2 (d100) for middle initial (Table #2): 83
  • Roll #3 (d100) for last name (Table #3): 48

“29” in table #2 gives: “Eric”. “83” in table #2 gives: “K.”. And “48” in table #3 gives: “Mitchell”. “Eric K. Mitchell”.

Here are another five names using this approach:

Roll #1 Roll #2 Roll #3 Resulting Name
22 26 57 Austin T. Cruz
64 92 46 Stephanie A. Rivera
25 74 68 Kevin D. Ortiz
53 99 78 Emily L. Cox
33 21 14 Samuel J. Wilson

(Table #4: Generated Names)

Now, let’s try something more advanced.

Generating Addresses

Let’s say you need to create a list of real-looking, but fake, street addresses. There are several parts that will need to be generated here. There’s a house or building number, a street name (which has several parts), there may be an apartment number, and finally a city, state, and zipcode. Let’s look at the tables that will help with this:

NOTE: Keep in mind that these are fictitious locations.

House or Building Number:
For the house number or building number, we’ll just directly use die rolls for that. We’ll roll a d20 and a d100 and concatenate the results. So, if you roll a “12” on the d20, and a “42” on the d100, the resulting house number will be “1242”.

Street Name:
Wikipedia has a helpful description of what components make up a street name:

Names are often given in a two-part form: an individual name known as the specific, and an indicator of the type of street, known as the generic. Examples are “Main Road”, “Fleet Street” and “Park Avenue”.

A street name can also include a direction (the cardinal points east, west, north, south, or the quadrants NW, NE, SW, SE), especially in cities with a grid-numbering system. Examples include “E Roosevelt Boulevard” and “14th Street NW”.

In the United States, most streets are named after numbers, landscapes, trees (a combination of trees and landscapes such as “Oakhill” is used often in residential areas), or the surname of an important individual (in some instances, it is just a commonly held surname such as Smith).

Based on the above, we’ll use the following tables to create a street name:

Include
Direction?
—roll d4
1. None 2. None 3. Prefix 4. Suffix

(Table #5: Street Name: Include Direction in Street Name?)

Direction
—roll d8
1. N 2. NE 3. E 4. SE
5. S 6. SW 7. W 8. NW

(Table #6: Street Name: Direction)

Specific
Name
Type
—roll d6
1. Cardinal
Directions		 2. Landmarks 3. Landscapes 4. Numbers 5. Presidents 6. Trees

(Table #7: Street Name: Identify Specific Name Type)

Once you identify the type of specific name, use the below table to pick the actual specific name:

Cardinal
Directions
—roll d4		 Landmarks
—roll d6		 Landscapes
—roll d12		 Numbers
—roll d12		 Presidents
—roll d8		 Trees
—roll d12
1. North 1. Broad 1. Creek 1. 1st 1. Adams 1. Cedar
2. East 2. Center 2. Forest 2. 2nd 2. Jackson 2. Cherry
3. South 3. Church 3. Highland 3. 3rd 3. Jefferson 3. Chestnut
4. West 4. High 4. Hill 4. 4th 4. Johnson 4. Dogwood
  5. Main 5. Lake 5. 5th 5. Lincoln 5. Elm
  6. Market 6. Lakeview 6. 6th 6. Madison 6. Hickory
    7. Meadow 7. 7th 7. Washington 7. Maple
    8. Orchard 8. 8th 8. Wilson 8. Oak
    9. Park 9. 9th   9. Pine
    10. Ridge 10. 10th   10. Spruce
    11. River 11. 11th   11. Walnut
    12. Sunset 12. 12th   12. Willow

(Table #8: Street Name: Select Actual Specific Name)

Generic
Name
—roll d12
1. Ave 2. Blvd 3. Cir 4. Ct 5. Dr 6. Ln
7. Pkwy 8. Pl 9. Rd 10. St 11. Ter 12. Way

(Table #9: Street Name: Generic Name)

Let’s look at an example street name using these tables:

  • Roll #1 (d4) for Including Direction (Table #5): 3
  • Roll #2 (d8) for Direction (Table #6): 6
  • Roll #3 (d6) for Type of Specific Name (Table #7): 5
  • Roll #4 (d8) for Specific Name (Table #8): 1
  • Roll #5 (d12) for Generic Name (Table #9): 9

Which gives: “SW Adams Rd”.

And another example street name:

  • Roll #1 (d4) for Including Direction (Table #5): 2
  • Roll #2 (d8) for Direction (Table #6): 0 (skipping)
  • Roll #3 (d6) for Type of Specific Name (Table #7): 6
  • Roll #4 (d12) for Specific Name (Table #8): 3
  • Roll #5 (d12) for Generic Name (Table #9): 5

Which gives: “Chestnut Dr”.

Apartment Number:
Roughly 1 in 8 people in the US live in apartments. We’ll use this ratio to decide if our address should include an apartment number.

Add
Apartment
Number?
—roll d8
1. No 2. No 3. No 4. No
5. No 6. No 7. No 8. Yes

(Table #10: Add Apartment Number?)

If we roll an “8” on a d8, then we’ll use a direct dice roll for the apartment number itself. A d100 will work here. So a roll of “72” would be “Apt 72” for address line 2.

City, State, Zip:
In the next three tables, you’ll need an unusual dice type, a d50, or 50-sided die. While such a die does exist, most people wouldn’t have one. They would either use a d100 and cut the result in half, or they would roll a d50 online, which you can do here.

For the city, we’ll use the following table of common city names:

City
—roll d50
1. Arlington 2. Ashland 3. Auburn 4. Bethel 5. Bloomington
6. Bristol 7. Brookview 8. Burlington 9. Centerville 10. Chester
11. Clayton 12. Cleveland 13. Clifton 14. Clinton 15. Dayton
16. Dover 17. Eden 18. Fairview 19. Farmington 20. Florence
21. Franklin 22. Georgetown 23. Glendale 24. Greenville 25. Greenwood
26. Hamilton 27. Hudson 28. Jackson 29. Kingston 30. Lebanon
31. Lexington 32. Liberty 33. Lincoln 34. Madison 35. Manchester
36. Marion 37. Midway 38. Milford 39. Milton 40. Mount Vernon
41. Newport 42. Oakland 43. Oxford 44. Pleasant Valley 45. Riverside
46. Salem 47. Springfield 48. Union 49. Washington 50. Winchester

(Table #11: City)

For the state, we’ll use the following table (leaving out territories for this example):

State
—roll d50
1. AL 2. AK 3. AZ 4. AR 5. CA
6. CO 7. CT 8. DE 9. FL 10. GA
11. HI 12. ID 13. IL 14. IN 15. IA
16. KS 17. KY 18. LA 19. ME 20. MD
21. MA 22. MI 23. MN 24. MS 25. MO
26. MT 27. NE 28. NV 29. NH 30. NJ
31. NM 32. NY 33. NC 34. ND 35. OH
36. OK 37. OR 38. PA 39. RI 40. SC
41. SD 42. TN 43. TX 44. UT 45. VT
46. VA 47. WA 48. WV 49. WI 50. WY

(Table #12: State)

And for zipcode, we’ll pull the first 3 digits of the zipcode from the following table using the same die roll as was used for the state. Then we’ll get the last 2 digits from another die roll (d100 to the rescue again) directly and pad with zero if necessary.

Zipcode
Prefix
—roll d50
1. 361 2. 998 3. 850 4. 722 5. 958
6. 802 7. 061 8. 199 9. 323 10. 303
11. 968 12. 837 13. 627 14. 462 15. 503
16. 666 17. 406 18. 708 19. 043 20. 214
21. 022 22. 489 23. 551 24. 830 25. 651
26. 596 27. 685 28. 897 29. 033 30. 086
31. 875 32. 122 33. 276 34. 585 35. 319
36. 731 37. 973 38. 171 39. 029 40. 292
41. 575 42. 372 43. 787 44. 841 45. 056
46. 232 47. 985 48. 253 49. 537 50. 820

(Table #13: Zipcode Prefixes)

Let’s look at an example city, state, zipcode using these tables:

  • Roll #1 (d50) for City (Table #11): 33
  • Roll #2 (d50) for State (Table #12): 29
  • Reuse Roll #2 (d50) for Zipcode Prefix (Table #13): 29
  • Roll #3 (d100) for remaining Zipcode digits: 7

Which gives: “Lincoln, NH 03307”.

And another example city, state, zipcode:

  • Roll #1 (d50) for City (Table #11): 42
  • Roll #2 (d50) for State (Table #12): 1
  • Reuse Roll #2 (d50) for Zipcode Prefix (Table #13): 1
  • Roll #3 (d100) for remaining Zipcode digits: 39

Which gives: “Oakland, AL 36139”.

Adding it all together:
Here are five full addresses using this approach:

House # Street Name Apt # City, State, Zip Generated Address
7, 41 1, 0, 2, 3, 11 3, 0 4, 32, 32, 36 741 Church Ter
Bethel, NY 12236
18, 89 3, 5, 6, 7, 9 6, 0 47, 2, 2, 54 1889 S Maple Rd
Springfield, AK 99854
11, 16 4, 3, 3, 5, 8 8, 26 19, 27, 27, 90 1116 Lake Pl E
Apt 26
Farmington, NE 68590
4, 13 2, 0, 5, 7, 2 8, 51 30, 13, 13, 05 413 Washington Blvd
Apt 51
Lebanon, IL 62705
5, 97 1, 0, 4, 3, 4 2, 0 44, 48, 48, 34 597 3rd Ct
Pleasant Valley, WV 25334

(Table #14: Generated Addresses)

Automating

Now, what if you needed to generate thousands or millions of these data points? Manually rolling dice won’t be convenient or practical. Some automation of the process will be needed. Luckily, computers do a pretty good job at producing random numbers and looking up data. Here’s what that process might look like:

  1. Write a script that produces a list of random numbers. For example, to produce a list of people’s names using tables 2 and 3, you would need three numbers between 1 and 100 for each name. Generate as many as are required for your purposes in a loop.

  2. Export the list of random numbers created in step 1.

  3. Load the needed lookup tables into a database, with a column holding the number you’re rolling for as the identifier of the row and another column holding the descriptive text.

  4. Import the list of random numbers from step 2 into your database.

  5. Write a query that joins the generated numbers with your lookup tables.

Keep in mind that you can extend or change your data tables however you see fit for your purposes. You can also come up with new data tables for scenarios we haven’t covered. And your random number generator can use values from any desired range; you’re not limited to just the values of physical dice. Just make sure that the range of the random number generated matches the number of records in your lookup table.

Enjoy!

<
Previous Post
Project Estimation
>
Next Post
C# and .NET News Updates