Barry Burgan
Activity
-
Barry Burgan replied to Nikhil Pandya
Death rate would be a continuous random variable rather than discrete (although you could do it categorically - (eg very low, low, medium, high, very high). Your analysis would look at infection and death rates in cities across the world (it is easy to get by country - https://www.worldometers.info/coronavirus/ but harder by city). That would give you an...
-
Barry Burgan replied to [Learner left FutureLearn]
These are great comments. Data analytics helps us understand more about a decision but does not make the decision for us - it is about applying the principles into a benefit cost models. So for example, one way to reduce the risk or spread of "skiable" days would be to buy a snow machine - but this comes at a cost, so that is what you need to use the data to...
-
Barry Burgan replied to Agnes T
Absolutely - a core assumption of finance is that investors (whether in share markets or in projects) are risk averse. This means they will only take on higher risk if there is an expected higher return. Different investors will be risk averse to different degrees.
-
Agree - excel will do the work for you, but it is helpful in this learning exercise to understand the basics
-
Some variables are determined by engineering parameters and so you can use probabilities (within a binomial etc) in that instance. Other variables are behavioural and require observation or judgement (an if - then type of analysis)
-
Yes you can calculate in in excel - will be based on the fit with a probability distribution and you can calculate from descriptive statistics if for example the distribution is normal. But excel also include options to calculate for a range of distributions
-
Not common in practice or examples, but yes - it depends on the nature of the variable. It is most common in change variables - like the change in items sold from one day to the next - the change in the number of televisions sold was -3
-
In the Insert option in the controls bar there is an option to insert equations (with symbols) or symbols directly
-
I agree Carlos - there is a lot of information even in the basic statistics. We tend to use averages/means a lot but don't delve into the information contained in spread (distribution, standard deviation) and shape (skewness) and yet when you interpret it as probabilities it tells us a lot about he risk of given decisions
-
Barry Burgan replied to Kyri Saphiris
Correct Kyri, on an assumption that the events are independent - well done
-
Hi Janet - the profit can be generated a number of ways - including doing market research and building an profit loss model, observing other operations, or using using historical data from previous leases (and maybe adjusting for changing circumstances). But you are right in that it does require some judgement
-
I agree that Excel allows you to do a lot of analysis, and it is so readily available, for general analysis. But there are a large number of other packages for even more extensive analysis (and as Nubia says, in some cases convenience of incorporating and merging data bases)
-
Thanks Fei Wang - your example is an engineering "based" source of probability. To extend this, business decisions are very commonly based on behavioural patterns, and as such require observation (frequency distribution) as the base
-
That's correct Jason
-
In this context Aprajita weight and probability are telling you the same thing. You are using the observed weights to assign probabilties
-
Good observation, but not necessarily 50-50% - it depends on the situation (eg in cards the probability of drawing a jack of spades from a full pack is 1/52, while the probability of drawing a jack is 4/52.
-
Hi Janet - hopefully the examples in the next topic (using excel to do the calculation will make it clearer)
-
The distribution is either engineering based (eg combinations of outcomes with dice) or more commonly in business decision making is observed as per the discussions on relative frequency in previous topic) - which is where big data comes in. Lots of data to choose from, and to break into sets. This is a very important issue, you need to be able to make...
-
The way to set it up in excel is covered in the next topic Linda
-
Barry Burgan replied to Janet Barnett
Correct - but an interesting issue is currency ranges and pricing strategies. A variable like height for example is actually continuous, but we tend to measure it like it is discrete (with lots of options of course) by saying someone is 176 cm when they could be 176.22216.
-
Barry Burgan replied to Kyri Saphiris
Thanks Kyri - that is clear (and what I expected). The issue is therefore you need a bit of extra known information to be able to close out the story completely if the events are not independent. Your story is tha they are not independent (otherwise P(A and B) = .26 (.4*.65). So if I presume that in your numbers that p(A and B but not C) = .25, and you add...
-
Barry Burgan replied to Kyri Saphiris
Hi Kyri - you have chosen a complex example - I don't think you have quite enough information to be able to calculate those seeking a tax advantage only. In your example (before I answer) can you let me know whether A,B and C are included in the A and B set?
-
Barry Burgan replied to Aayushmaan Chauhan
Thanks for the reply Aprajita - and a good question Aasyushmaan. How you use the principles will vary based on your data set and your situation (question) but the principles remain! We make decisions in a world of uncertainty and analysing the data (and interpreting the issue in a probability context) helps us understand the risks
-
These a great questions and a a good reply. You have the probability distribution overall, but you can look at subsets of the data to get probabilities for that subset (so you could look at the subsets of attendance at training in morning, afternoon and evening sessions). The deeper you delve into the data more you can inform the decisions you will dtake
-
Agree, thanks for the suggestion. The book has great examples of applications in decision making
-
That is correct Faisal - probability means relative frequency, and is used in situations where you can observe that relative frequency. Of course one of the interesting issues in business decision making is how you predict relative frequencies into the future. You can observe the past, but you then need to allow for environmental changes and how it impacts...
-
Significance and confidence levels apply most directly to the reliability (interpreted in terms of probability) with respect to sample parameters - that is, how reliable are they as estimates of the population parameter
-
Barry Burgan replied to Oswaldo velez-langs
Good comment Oswaldo - a hard issue is to determine how many bins to break nominal data into, so that you can see the patterns
-
A very good comment Michael. The graph needs to be effective in telling the story. This requires judgement from the Excel users. I also agree that presenting the data badly can cause "disinformation" and some people actually do that deliberately. As a quick summary to me the "best" type of graph is linked to the nature of the data. Pie charts are commonly...
-
Hi Juan - I responded to an earlier comment in this. The "best" type of graph is linked to the nature of the data. Pie charts are commonly used to look a the distribution of categorical data (or grouped continuous data) and are visually appealing. Bar charts and histograms are more useful in this context if you are trying to visualise a probability...
-
Hi - the best type of graph used is often dependent on the nature of the data being examined. Some graphical procedures are better for some data than others. My experience is that the three most common types of graph used are line plots (especially for time linked data, but also for when trying to identify correlations), pi-charts when looking at spread of...
-
Hi - can you let me know which lecture specifically you can't play. There is no video for 1.5, it is an introductory comment. The videos are for 1.6 etc