Business Analytics

Market Intelligence | Data Analytics

[Business Analytics @_Market_Intel]

@JDave13579 Correlation gives us an idea about the strength (how close to 1 or -1) and direction (positive or negative) of the relationship. Regression explains the average movement in a Y (dependent) variable based on changes in the independent variable(s).

[Business Analytics @_Market_Intel]

@TeaBatte Regression helps us understand how much (on average) one variable changes based on changes in another variable.

[Business Analytics @_Market_Intel]

@JoshuaScarbro13 Key distinction: with ANOVA, we can learn which fertilizers are related to larger crop yields, but we cannot determine whether those fertilizers caused the larger yield. For that we would need to conduct a larger/more complicated experiment.

[Business Analytics @_Market_Intel]

@MeaganR03957376 ANOVA helps determine if differences exist between groups, but cannot be used to determine whether one variable is causing a change in another variable.

[Business Analytics @_Market_Intel]

@bullock_braxton Would suggest doing each variable seperately, then with both (with interaction). We could use some other metrics to compare and then choose the best model.

[Business Analytics @_Market_Intel]

This is an interesting research question. We would need to propose an expected value. Would we expect the college graduation rate to be the same, higher, or lower for 10 and 25 year-later students....than for those just out of high school?

[Lisa Morgan @LisaMorg9816]

@_Market_Intel Propose using Chi-Square method to analysis the graduation rate of students attending college at different phases in their lives. After immediately high school, 10 and 25 years into their careers.

[Business Analytics @_Market_Intel]

Great tweet below. Question: Can Chi-Square be used to test this assumption? If so, how?

Advertising Effectiveness: "Younger use Tiktok/Instagram and older use Facebook/Email."

[Meagan Herring @MeaganR03957376]

@_Market_Intel
Social media platforms are used to reach different generations of people. Younger are tiktok/Instagram and older Facebook/email. Businesses have to know where is the best place to advertise to their target consumer.

[Business Analytics @_Market_Intel]

Somebody call Wash DC, we need to get the government involved.

https://t.co/Jtmj0PoW6z

[Business Analytics @_Market_Intel]

Great headline! An analytical review of their earnings reminds us that net income doesn’t always reflect quarterly operations. But you gotta love how Tesla seemingly always finds a way to deliver. https://t.co/KPGt7BMBhV

[Business Analytics @_Market_Intel]

@LeahTolbert12 Thanks for the tweet Leah. Quick clarification: statistical significance refers to very low probabilities (p-value) of committing Type I error when rejecting the Null. We typically accept the alternative when p-value is "highly significant" which is often considered (p < .05).

[Business Analytics @_Market_Intel]

You know it’s going to be good when the sandwich insurance is only $0.15. 😎

https://t.co/JR2n37jxKK

[Business Analytics @_Market_Intel]

@JDave13579 Hang in there Justin! Trying to build a solid foundation for everyone...before things start getting complicated.

[Business Analytics @_Market_Intel]

@CJHarri94128778 CJ great question. Regressions always have a constant. With dummy variables, you essentially set each variable category as another "constant" that is multiplied by the regression coefficient. Dummy variables are ordinal data, thus interpreted differently than interval/ratio data.

[Business Analytics @_Market_Intel]

@Cameron25446968 @EnglandKile @SamHoov53011989 This study provides evidence that playing Grand Theft Auto V does not lead to higher levels of "aggression, empathy, interpersonal competencies, impulsivity-related constructs, depressivion, anxiety or executive control functions."

[Business Analytics @_Market_Intel]

@SamHoov53011989 @EnglandKile thinks maybe this correlation will not be statistically significant.

[Business Analytics @_Market_Intel]

@CJHarri94128778 Yes. More sophisticated statistical packages allow you to choose how (and in what order) each variable is inserted into the equation. Two advanced procedures for multivariate regression include stepwise analysis and evaluation of multicollinearity.

[Business Analytics @_Market_Intel]

@kristakathleen_ ANOVA can distinguish between groups - but correlation (and preferably regression) are necessary to evaluate whether changes in one variable are related to changes in another.

[Business Analytics @_Market_Intel]

@CJHarri94128778 Post-hoc tests are useful for determining differences among groups for more than two variables (with ANOVA).

[Business Analytics @_Market_Intel]

Oh we are live again @USA_MCOB this fall! You know you want to follow this @montrezradcliff ...stats knowledge might improve your game...

[Devin Curry @devin_curry10]

@reece_bray @_Market_Intel @montrezradcliff Im gonna bring it back. And I’m not even in school any more. 😂 Let me find a book.

[Business Analytics @_Market_Intel]

Shout to ⁦@ghendrix14⁩ for your business-academic success. Well done! https://t.co/Hq9PcmGQ64