Hi All,(adsbygoogle = window.adsbygoogle || []).push({});

Hope this is the right sub-forum for this.

I currently work for an engineering consultancy and am doing some project management what for what are essentially site investigation reports (sites are quite small, typically a few acres/hectares). Our clients are mostly developers who have purchased land and want to build on it (can be residential, commercial or industrial development, it depends on the client).

I need to put together cost quotes for the site investigation work. I've not been asked to do this before and so was looking at how it's been done in the past and it seems to be mostly based on people's past experience/judgement calls etc. However, I want to put us on a more scientific footing rather than trying to "guess" what the answer will be. I was therefore considering multivariate regression analysis to try and determine what the contributing variables are to the actual cost and to come up with a mathematical model to try and predict future costs based on a limited range of site characteristics.

We have cost data on previous projects, principally the estimated cost, the hours staff spent on it, and then the actual cost (based on the actual staff hours). I also have information relating to each site that was investigated. Some of this I assume would be classed as "quantitative" data such as size of site (in hectares). However, a lot of it is what I would call "qualitative" such as as initial assessment of the level of pollution likely to be on the site, e.g. this is a simple done on a simple scale of 1-3, 1 = low, 2 = moderate, 3 = high.

One of the purposes of the site investigation is to quantify the level of pollution better but at the start we just have a broadscale (essentially desktop study) estimate based on the 1, 2 or 3 scale.

My question is therefore: can I use these sorts of scales in a multivariate regression analysis? In the above example the range 1, 2 or 3 (which is a qualitative judgement), I am not sure that this is treated the same as, for example, the site area (in hectares), which is a quantitative measure where I can simply compare one data point from the site equally with all the others.

Basically I need advice on how to take qualitative values and include them in multivariate regression analysis, as I am not even sure you are supposed to do this, let alone how to do it.

Many thanks in advance for any help provided. :)

Regards

Richard

Ps - I intend to use Excel for the analysis. I have tutorials which show how to use the LINEST function for this sort of thing. I believe I can do this part of the analysis, it's just how to classify/input the data into the model that I am unsure about.

**Physics Forums | Science Articles, Homework Help, Discussion**

The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

# Multivariate regression for project cost estimates

**Physics Forums | Science Articles, Homework Help, Discussion**