Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Dairy farming algorithms for decision optimization?

Hi,

Does anyone have an interest in developing some algorithms for automatically running decision optimization or quantitative management? This isn't really a trading algo, but I'm looking for someone interested to work with our team.

We're building a command center to handle incoming data from hundreds of dairy farms. Feed data, milk output data, lifespan productivity, calf sales, beef sales, milk sales, itemized expenses... everything. Does anyone here know how to work within ruby on rails (its like python) to build an algo to identify the best and worst correlations within the data, so that we can implement appropriate changes across multiple farms? For example, if an algo could make management recommendations such as: "Farms with aerated manure lagoons report 40% less solids and $150,000 less in waste handling expenses on average." That's the sort of data correlation we can apply right away to numerous other farms.

We are anticipating massive amounts of data and are seeking some automated manner to analyze and quantify significance. If you are interested to get involved and be a part of this company, please contact me. The dairy industry has a lot of room for improvement via the application of technology.

Thanks,
Taylor

2 responses

Hello Mr. Moffitt,
I certainly agree, the dairy industry does have much room for improvement but it is far more advanced than the general public realizes. The industry is often on leading the way in various innovative sectors such as reproduction (modern human artificial reproduction owes much to pioneering research from the dairy sector), food preservation and robotics to name a few.

Pertaining specifically to aerated manure lagoons, these are usually installed to reduce odor problems, not to reduce costs, as the aeration process has associated operating costs. Here is a good factsheet on costs: http://www.omafra.gov.on.ca/english/engineer/facts/04-033.htm If you are indeed having odor problems, you might find it easier to have public engagement with those retirees/halfbacks. Now that they have moved to a rural area from an urban one, explain to them the importance of locally grown food. Also point out the country is where country business takes place, if they want a conveniently controlled city life, they can find that in the city. I’m sure you can find some more diplomatic words. The aeration process releases greenhouse gases and as such may not be as attractive in years to come. Through that process there is also a substantial loss of Nitrogen. Depending on your farms distance to crop fields it is generally more cost effective to conserve that manure N for field application in order to reduce artificial fertilizer costs (https://www.extension.iastate.edu/agdm/livestock/html/b1-65.html). If indeed trucking costs are an issue, you might look at manure solids as bedding. If N application is not necessary then you might also look into generating electricity from a manure digester. That would mostly depend on incentives available in your area for renewable energy production. Iowa State University might have some of the numbers you are looking for here: https://www.extension.iastate.edu/agdm/ldcostsreturns.html

I think what you are trying to do is interesting. Im not a coder by profession, I taught myself Python, but I do have some ideas I think might be of use to you. These are based on my own experience having grown up on a dairy farm that just installed 2 robotic LELY milkers (https://www.youtube.com/watch?v=Z7ibgGPTOUQ) in spring 2016 and based on my experience as an Animal Nutritionist with a Masters in Animal Nutrition.

  1. Install as many robotic milkers as fast as possible. Humans are error prone and can sometimes be non-reliable, especially on Sunday morning (after a night out). Get rid of that uncertainty by installing robots. The ID collars can be used for heat detection as well. The farms near the area I grew up have seen 5 to 25% increases in milk production since installing the robots. I think this was mostly due to increased milkings/d as most were milking only 2x. Feed costs have also been reduced since they are now individually fed as opposed to TMR (reducing fuel costs). The system will also generate considerable amount of individual data for each cow which then be used to adjust herd practices accordingly.

  2. Install acidified calf milk feeders or robotic calf feeders (like those offered by URBAN: https://www.youtube.com/watch?v=OBkzHhysMDU) for calves. This will increase calf growth potential and reduce substantially labor costs.

  3. Encourage local farmers to share equipment (https://www.washingtonpost.com/news/the-switch/wp/2016/05/06/meet-the-site-that-is-like-uber-but-for-tractors/?utm_term=.65cbcb03231d). Depending on the size of farms, they often have way too much machinery. Farmers like to be independent and have their own piece of machinery available when they want it, but there are far too many random pieces depreciating/rusting away because of a lack of communication and teamwork. They could benefit from economies of scale as well.

  4. Make use of an animal nutritionist. There are many by-products that can be implemented in feed rations, thus lowering feed expenses. When I was working as a nutritionist I was able to reduce farm feed costs by using products such as: corn distillers (by-product of ethanol), waste potatoes, waste french fries, waste oil, waste candy (like Jelly Beans) and waste carrots to name a few.

These are the 4 major areas in my opinion where currently the most tangible increases to revenue and decreases to expenses can be found on most dairy farms based on my experiences in the dairy industry. I would also suggest that some of the topics might be useful to graduate students as Universities are always looking to engage with businesses for the purpose of doing research in practical areas.

There is so much more that could be said on each topic. If you need further information don`t hesitate to contact me.

Have a good day.

Having some background about the science behind dairy farming, I find these problems interesting. In general there are specialists who can consult with dairy farmers, and they mostly rely on published research. In the example with "aerated manure lagoons", there would likely be multiple studies discussing the impact and the peer review process makes sure the authors are not fooling themselves through bad data analysis. For example, alternative explanations have to be ruled out. Thus is the work of science...

In a sense, what you are asking for is automated data mining, and there really is no such thing. The first step always is to analyze the data by hand. Then you can prepare a report to make a case with the decision makers, and hope they believe you and you are right. Sometimes you can even make the cases for an automated decision making process, much like a trading algo.

There is no automated process for science, no matter what some salespeople say. You always have to think about the analysis and the conclusions. You need somebody who knows data analysis, probably a data scientist. Then there is a whole ton of regulation with what kind of data you are allowed to share, especially if you don't own all the farms.

And you are not going to keep Ruby for the data analysis part, I can tell you that much. Python and R are way more entrenched in data analysis, and doing stuff like that in Ruby will be a daily uphill battle. Ultimately, of course, it is possible to do all sorts of data analysis in Ruby, but you'll have to research harder and write a lot more code on your own rather than use stackoverflow or nice packages. Also, data analysis is almost never done in a web server process, but separately, often in a regular interval like once a day or an hour, and the results distributed through a database.