Chili for a Chilly Day

I haven't posted a recipe in a while, but as Friday swung our weather from bright blue, warm days into chilly, rainy winter in a quick snap of the fingers, I thought it was time to make the first chili of the year. I've been building this chili recipe since high school, when I first added a block of unsweetened chocolate to the mix, into college when I first added dark beer. Now, the recipe contains hints of a molé sauce as well, and I make my spice mix in advance, to allow the flavours to blend. Oh, and open a bottle of your favorite dark beer, or two if you want to start drinking B) Set the beer aside to become flat.


While adding beans are optional, I'm planning to do so, and since I'll be using dried beans. This step has the longest lead time. First, I buy my dried beans at Phipps Country Store and Farm in Pescadero, CA. They are about an half-hour drive from me, and I'll visit them several times a year to replenish my supply of dried beans. They have an huge selection of dried beans. For red chili, I use a combination of black beans and one or more dried beans from the kidney family: Big Mexican Red Kidney, Cranberry, Pinto or Red beans. Cranberry are my favorite; they're a big, meaty bean, with a nutty flavour that compliments the creamy black bean nicely. Phipps now has an online store, so you can buy their great beans even if they aren't a convenient drive from you.

I'll use about two pounds of beans, one pound of dried black beans, and the second pound made up of whatever kidney varietals I'm using. As I said, Cranberry beans are my favorite for chili, and that's while I'll be using with the black beans today. Put the dried beans in a strainer, and rinse under cold water. Carefully check the beans, removing any discolored, withered or soft beans, as well as any foreign material such as stems or stones. Place the beans in a kettle and cover with enough cold water to top the beans by two inches. Remove any "floaters". Add a bay leaf. Do not add any salt or acids [tomato, vinegar, etc] as these will wrinkle the beans. You can let the beans soak overnight, or bring the kettle to a boil, simmer the beans for five minutes, and then let the beans soak in the hot water, covered, for an hour. After the hour soak, remove the beans, retaining about a cup of the water and the bay leaf. All of this just prepares the beans. They're not cooked and ready to eat yet.

Prepared Beans in a Ceramic Bowl

Either in advance, or an hour before serving, put the prepared beans back into the kettle, add the reserved soaking liquor and bay leaf and a red [hot] or yellow [sweet] onion, peeled and studded with cloves. Do not add salt nor acids. Cover with enough cold water to just top the beans. Bring to a boil, place on simmering bricks, and simmer for 45 minutes or until the beans are tender.

Chili Base

This is the real "Chili" with Tex-Mex, "Texas Red", Chili con Carne, and Chili with Beans being stews based upon Chili. I start with about five pounds of tomatoes and five pounds of peppers. The tomatoes can be heirloom, cluster, or whatever you have in your garden or local store that are fresh, feel heavy for their size and are very ripe. If you use anything other than red tomatoes, your chili may have an odd colour, but the flavour will be great. I usually use an equal, by weight, combination of chili peppers and bell peppers. In California, at this time of year, there is a great selection of chilies: poblano, anaheim, astor, etc. I generally avoid green bell peppers, as I prefer the flavour of the red, orange and yellow ones. Today I'm using almost three pounds of poblano chilies and two pounds of red, orange and yellow bell peppers.

I start with the tomatoes, as they'll take awhile as well, and can also be prepared the day before, as the beans can. Bring a large pot or kettle of cold, salted water to a boil. While it's coming to a boil, using a very sharp knife, I use a "bird's beak" hooked knife, remove the stem end from each tomato, and score an "X" in the skin at the opposite end from the stem. Place the tomatoes into the boiling water. Once the pot has returned to a boil, after two minutes or so, the skin will begin to peel back from the scored end of the tomatoes. Remove the tomatoes, and place in a bowl to cool. As soon as you can handle them, remove the skin from the tomatoes. Cut each tomato in half, cross-wise, and remove the seeds with a small spoon. Place the peeled, cored tomatoes, cut-side down, into a colander, and allow to drain for at least an hour, but overnight in the refrigerator [over a bowl] is fine too. You'll be amazed how much water you'll collect. You can save this tomato juice, to use in place of water in stock or stews, or to thin this chili, if needed.

Peeled Tomatoes
Tomatoes Cut Along the Cross Section
Tomatoes With Seeds Removed

Next, rinse the peppers and fire-roast them, either over the flames on a gas-stove, under a broiler, or over a grill. Leave the peppers whole, and flame them until the skin is blackened all over each pepper. Place the peppers into a paper bag, or wrap in parchment paper - this traps enough steam to help loosen the skins, and allow them to cool. If the peppers are hot to the taste [such as an jalapeño chili], wear rubber gloves and a mask to avoid capsicum burns. Scrape as much skin as possible off of the peppers with a knife, core them, cut in half, lengthwise, and remove the white veins. Cut the peppers into strips, lengthwise.

Peppers Being Fire Roasted over a Gas Stove Top

Make a soffritto of one sliced large red onion, two crushed cloves of garlic, the peppers, your favorite chili powder and cilantro leaves that have been rinsed, dried and chopped. A soffritto is just a slowly cooked medley of vegetables, spices and herbs in olive oil. After an hour or two or three, chop those tomatoes that have been draining and add those and that bottle of beer that you were leaving to become flat, a block of very good dark, unsweetened chocolate, and two tablespoons of freshly ground, roasted, unsalted valencia [sweet, and what I prefer] or virginia [meatier tasting] peanuts. Stir it around, add a teaspoon each of coarse sea salt, Mexican oregano and cayenne pepper, and cook on the simmering bricks or in an oven on low, for about an hour. Add salt and seasonings to taste.

Drained Tomatoes on the Soffritto
Block of Dark Chocolate on a Plate
Chocolate Added to the Soffritto
Ground Peanuts added to the Soffritto
Chili Simmering in the Pot

Chili con Carne

If you're going to make a chili con carne, take two pounds of cubed beef, and brown the cubes in bacon fat. Add the chili over the meat to cover, and let simmer another hour. Shred the meat cubes apart using two forks and return to the stew, or slice thinly. Add the cooked beans and serve.

Chili Con Carne
Chili Con Carne with Beans in the Pan

Vegetarian Chili Frijole

Add the cooked beans to about a quart of the chili and serve.

Serving Suggestions

  1. Put a cooked, hot tamale, of your favorite variation into a bowl and cover with the chili. I like pork tamale with the chili con carne, and chilies & cheese tamale with the vegetarian chili
  2. Put cooked, brown, short-grain rice into a bowl, top with chili.
  3. Fill a bowl with chili and serve with hot cornbread. I especially like the recipe from Recipes for Living in Big Sur. This corn bread is stuffed with chilies and corn, and layered with cheese.
  4. Enjoy it diner style, served in a bowl with oyster or soda crackers.
  5. Forget the beans, don't shred the beef, and enjoy a bowl of Texas Red.
  6. Put out various condiments: chopped onions, chopped, fire-roasted jalapeños or hotter chilies, hot pepper sauces [there are many on the market, try something new], shredded cheese - especially Mexican cheeses.
  7. Serve over a bowl of different grains, rather than rice: quinoa is a great choice, especially for the vegetarian version, soft polenta is another good choice, and you can check out many more suitable grains at Bob's Read Mill.

  8. Take a crunchy deli roll, hollow it out, fill with thinly sliced beef and top with the chili and cheese. Or forget the beef, and use a kielbasa sausage.
ttp://">Chili Served over a Corn Muffin in flat Bowl


First and foremost, recipes are guidelines, not exact instructions that you must follow. Add more or less of anything. Consider every recipe a starting point for your own imagination and taste.

We already talked about the various beans you can use, as well as the variety of chilies. There are many more types, of course. Be adventurous.

Instead of beef, try other red or dark meats: venison, buffalo or beefalo, elk, duck, turkey thighs - especially from a wild turkey, or other game meat. Go wild.

Rather than cayenne pepper, use ground ancho [sweet and fruity] or chipotle [smoky] chilies.

Try chili verde. Use tomatillos rather than tomatoes, forget the chocolate and peanut butter. Use white beans rather than red, and white meats rather than red.

Chili Powder

I make my own chili powder. I start by filling an old spice jar, 50/50 with cumin seeds and coriander, and shoving in a cinnamon stick. When I need chili powder, I take a teaspoon of the mixture, and toast the seeds. Allow the cumin and coriander to cool, then grind in a mortar and pestle, add a 1/4 teaspoon of ground cinnamon and a 1/2 teaspoon of ground allspice.


Reading Pentaho Kettle Solutions

On a rainy day, there's nothing better than to be sitting by the stove, stirring a big kettle with a finely turned spoon. I might be cooking up a nice meal of Abruzzo Maccheroni alla Chitarra con Polpettine, but actually, I'm reading the ebook edition of Pentaho Kettle Solutions: Building Open Source ETL Solutions with Pentaho Data Integration on my iPhone.

Some of my notes made while reading Pentaho Kettle Solutinos:

…45% of all ETL is still done by hand-coded programs/scripts… made sense when… tools have 6-figure price tags… Actually, some extractions and many transformations can't be done natively in high-priced tools like Informatica and Ab Initio.

Jobs, transformations, steps and hops are the basic building blocks of KETTLE processes

It's great to see the Agile Manisto quoted at the beginning of the discussion of AgileBI. 

BayAreaUseR October Special Event

Zhou Yu organized a great special event for the San Francisco Bay Area Use R group, and has asked me to post the slide decks for download. Here they are:

No longer missing is the very interesting presentation by Yasemin Atalay showing the difference in plotting analysis using the Windermere Humic Aqueous Model for river water environmental factors, without using R and then the increased in variety and accuracy of analysis and plotting gained by using R.

Search Terms for Data Management & Analytics

Recently, for a prospective customer, I created a list of some search terms to provide them with some "late night" reading on data management & analytics. I've tried these terms out on Google, and as suspected, for most, the first hit is for Wikipedia. While most articles in Wikipedia need to be taken with a grain of salt, they will give you a good overview. [By the way, I use the "Talk" page on the articles to see the discussion and arguments about the article's content as an indicator of how big a grain of salt is needed for that article] &#59;) So plug these into your favorite search engine, and happy reading.

  • Reporting - top two hits on Google are Wikipedia, and, interestingly, Pentaho
  • Ad-hoc reporting
  • OLAP - one of the first page hits is for Julian Hyde's blog, creator of the open source tool for OLAP, Mondrian, as well as real-time analytics engine, SQLstream
  • Enterprise dashboard - interestingly, Wikipedia doesn't come up in the top hits for this term on Google, so here's a link for Wikipedia:
  • Analytics - isn't very useful as a search term, but the product page from SAS gives a nice overview
  • Advanced Analytics - is mostly marketing buzz, so be wary of anything that you find using this as search term

Often, Data Mining, Machine Learning and Predictives are used interchangeably. This isn't really correct, as you can see from the following five search terms…

  • Data Mining
  • Machine Learning
  • Predictive Analytics
  • Predictive Intelligence - is an earlier term for Predictives that has mostly been supplanted by Predictive Analytics. I actually prefer just "Predictives".
  • PMML - Predictive Modeling Markup Language - is a way of transporting predictive models from one software package to another. Few packages will both export and import PMML. The lack of that capability can lock you into a solution, making it expensive to change vendors. The first hit for PMML on Google today is the Data Mining Group, which is a great resource. One company listed, Zementis, is a start-up that is becoming a leader in running data mining and predictive models that have been created anywhere
  • R - the R statistical language, is difficult to search on Google. Go to and … instead. R is useful for writing applications for any type of statistical analysis, and is invaluable for creating new algorithms and predictive models
  • ETL - Extract, Transform & Load, is the most common way of getting information from source systems to analytic systems
  • ReSTful Web Services - Representational State Transfer - can expose data as a web service using the four verbs of the web
  • SOA
  • ADBMS - Analytic Database Management Systems doesn't work well as a search term. Start with the site and follow the links from the Eigenbase subproject, LucidDB. Also, check out AsterData
  • Bayes - The Reverend Thomas Bayes came up with this interesting approach to statistical analysis in the 1700s. I first started creating Bayesian statistical methods and algorithms for predicting reliability and risk associated with solid propellant rockets. You'll find good articles using Bayes as a search term in Google. A bit denser article can be found at And some interesting research using Bayes can be found at: Andrew Gelman's Blog. You're likely familiar with one common Bayesian algorithm, naïve Bayes, which is used by most anti-spam email programs. Other forms are objective Bayes with non-informative priors and the original Subjective Bayes. I have an old aerospace joke about the Rand Corporation's Delphi method, based on subjective Bayes :-) I created my own methodology, and don't really care for naïve Bayes nor non-informative priors.
  • Sentiment Analysis - which is one of Seth Grimes' current areas of research
  • Decision Support Systems - in addition to searching on Google, you might find my recent OSS DSS Study Guide of interest

Let me know if I missed your favorite search term for data management & analytics.

Data Artisan Smith or Scientist

Over the past few months, a debate has been proceeding on whether or not a new discipline, a new career path, is emerging from the tsunami of data bearing down on us. The need for a new type of Renaissance [Wo]Man to deal with the Big Data onslaught. To whit, Data Science.

I'm writing about this now, because last night, at an every-three-week get together devoted to cask beer and data analysis, the topic came up. [Yes, every-THREE-weeks - a month is too long to go without cask beer fueled discussions of Rstats, BigData, Streaming SQL, BI and more.] The statisticians in the group, including myself, strongly disagreed with the way the term is being used; the software/database types were either in favor or ambivalent. We all agreed that a new, interdisciplinary approach to Big Data is needed. Oh, and I'll stay on topic here, and not get into another debate as to the definition of "Big Data". &#59;)

This lively conversation reinforced my desire to write about Data Science that swelled up in me after reading "What is Data Science?" by Mike Loukides published on O'Reilly Radar, and a subsequent discussion on Twitter held the following weekend, concerning data analytics.

The term "Data Science" isn't new, but it is taking on new meanings. The Journal of Data Science published JDS volume 1, issue 1 in January of 2003. The Scope of the JDS is very clearly related to applied statistics

By "Data Science", we mean almost everything that has something to do with data: Collecting, analyzing, modeling...... yet the most important part is its applications --- all sorts of applications. This journal is devoted to applications of statistical methods at large.
-- About JDS, Scope, First Paragraph

There is also the CODATA Data Science Journal, which appears to have last been updated in August of 2007, and currently has no content, other than its self-description as

The Data Science Journal is a peer-reviewed electronic journal publishing papers on the management of data and databases in Science and Technology.

I think that two definitions can be derived from these two journals.

  1. Data Science is systematic study, through observation and experiment, of the collection, modeling, analysis, visualization, dissemination, and application of data.
  2. Data Science is the use of data and database technology within physical and natural sciences and engineering.

I can agree with the first, especially with the JDS Scope clearly stating that Data Science is applied statistics.

The New Oxford American Dictionary, on which the Apple Dictionary program is based, defines science as a noun

the intellectual and practical activity encompassing the systematic study of the structure and behaviour of the physical and natural world through observations and experiments.

And a similar definition of science can be found on

In many ways, I like Mike Loukides' article "What is Data Science?" in how it highlights the need for this new discipline. I just don't like what he describes to be the new definition of "data science". Indeed, I very much disagree with this statement from the article.

Using data effectively requires something different from traditional statistics, where actuaries in business suits perform arcane but fairly well-defined kinds of analysis. What differentiates data science from statistics is that data science is a holistic approach. We're increasingly finding data in the wild, and data scientists are involved with gathering data, massaging it into a tractable form, making it tell its story, and presenting that story to others.

A statistician is not an actuary. They're very different roles. I know this because I worked for over a decade applying statistics to determining the reliability and risk associated with very large, complex systems such as rockets and space-borne astrophysics observatories. I once hired a Cal student as an intern because she feared that the only career open to her as a math major, was to be an actuary. I showed her a different path. So, yes, I know, from experience, that a statistician is not an actuary. Actually, the definition of a data scientist given, that is "gathering data, massaging it into a tractable form, making it tell its story, and presenting that story to others" is exactly what a statistician does.

I do however see the need for a new discipline, separate from applied statistics, or data science. The massive amount of data to come from an instrumented world with strongly interconnected people and machines, and real-time analysis, inference and prediction from those data, will require inter-disciplinary skills. But I see those skills coming together in a person who is more of a smith, or, as Julian Hyde put it last night, an artisan. Falling back on the old dictionary again, a smith is someone who is skilled in creating something with a specific material; an artisan is someone who is skilled in a craft, making things by hand.

Another reason that I don't like the term "data science" for this interdisciplinary role, stems from what Mike Loukides describes in his article "What is Data Science?" as the definition for this new discipline "Data science requires skills ranging from traditional computer science to mathematics to art". I agree that the new discipline requires these three things, and more, even softer skills. I disagree that these add up to data science.

I even prefer "data geek", as defined by Michael E. Driscoll in "The Three Sexy Skills of Data Geeks". Michael Driscoll's post of 2009 May 27 certainly agrees skill-wise with Mike Loukides post of 2010 June 02.

  1. Skill #1: Statistics (Studying)
  2. Skill #2: Data Munging (Suffering)
  3. Skill #3: Visualization (Storytelling)

And I very much prefer "Data Munging" to "Computer Science" as one of the three skills.

I'll stick to the definition that I gave above for data science as "systematic study, through observation and experiment, of the collection, modeling, analysis, visualization, dissemination, and application of data". This is also applied statistics. So, what else is needed for this new discipline? Well, Mike and Michael are correct: computer skills, especially data munging, and art. Well, any statistician today has computer skills, generally in one or more of SAS, SPSS, R, S-plus, Python, SQL, Stata, MatLab and other software packages, as well as familiarity with various data storage & management methods. Some statisticians are even artists, perhaps as story tellers, as evidenced by that rare great teacher or convincing expert witness, perhaps as visualizers, creating statistically accurate animations to clearly describe the analysis, as evidenced by the career of that intern I hired so many years ago.

The data smith, the data artisan, must be comfortable with all forms of data:

  • structured,
  • unstructured and
  • semi-structured

Just as any other smith, someone following this new discipline might serve an apprenticeship creating new things from these forms of data such as a data warehouse or an OLAP cube, a sentiment analysis or a streaming SQL sensor web, or a recommendation engine or complex system predictives. The data smith must become very comfortable with putting all forms of data together in new ways, to come to new conclusions.

Just as a goldsmith will never make a piece of jewelry identical to the one finished days before, just as art can be forged but not duplicated, the data smith, the data artisan will glean new inferences every time they look at the data, will make new predictions with every new datum, and the story they tell, the picture they paint, will be different each time.

And perhaps then, the data smith becomes a master, an artisan.

PS: Here's a list of links to that Twitter conversation among some of the most respected people in the biz, on Data Analytics


December 2019
Mon Tue Wed Thu Fri Sat Sun
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          
 << <   > >>
The TeleInterActive Press is a collection of blogs by Clarise Z. Doval Santos and Joseph A. di Paolantonio, covering the Internet of Things, Data Management and Analytics, and other topics for business and pleasure. 37.540686772871 -122.516149406889



The TeleInterActive Lifestyle

Yackity Blog Blog

The Cynosural Blog

Open Source Solutions


The TeleInterActive Press

  XML Feeds


Our current thinking on sensor analytics ecosystems (SAE) bringing together critical solution spaces best addressed by Internet of Things (IoT) and advances in Data Management and Analytics (DMA) is updated frequently. The following links to a static, scaleable vector graphic of the mindmap.

Recent Posts

free blog software