Wrote these queries as example queries on sensor data - http://rethinkdb.com/blog/temperature-sensors-and-a-side-of-pi/

rethinkdb_statistical_reql_query_examples.py

#!/usr/bin/python
import rethinkdb as r
from datetime import datetime, timedelta
conn = r.connect("localhost", 28015, db="telemetry_pi")

##
#Finding the average temperature & humidity for the past 24 hours
##
day_ago = datetime.now() - timedelta(hours=24)
cursor = r.table("observations")\
  .filter(r.row("datetime") > day_ago))\
  "avg_humidity": r.avg(r.row("temp")),
  .merge({
    "avg_temperature": r.avg(r.row("humidity"))
    }).run(conn)

##
#Finding the hottest observations per day (or a _single_ maxima for each day)
##
cursor = r.table("observations").group(
  lambda doc: return [
      doc('datetime').year(),
      doc('datetime').month(),
      doc('datetime').day()
  ])\
  .max("temp")\
  .ungroup()\
  ["reduction"]

##
#Finding some basic statistics and calculating a simple linear regression (y_humidity = alpha + beta_temp * temp)  
##
# https://en.wikipedia.org/wiki/Simple_linear_regression
# https://en.wikipedia.org/wiki/Pearson_product-moment_correlation_coefficient
##

#To dramatically shorten queries
obs = r.db("telemetry_pi").table("observations")

#Start with initial object containing calculations that don't have dependencies
r.object(

  #Sums
  "sum_humidity", obs.sum("humidity"),
  "sum_temp", obs.sum("temp"),

  #Averages
  "avg_humidity", obs.avg("humidity"),
  "avg_temp", obs.avg("temp"),

  #Sum of Squares
  "sumsq_humidity",obs["humidity"].map(lambda d: return d * d).sum(),
  "sumsq_temp", obs["temp"].map(lambda d: return d * d ).sum(),

  "sum_products_temp_humidity", r.map(
      obs["temp"],
      obs["humidity"],
      lambda temp, humidity:
      	return temp * humidity
    ).sum(),

  #n (for convenience)
  "n", obs.count()

#Now calculate beta estimator and standard deviations and Pearson correlation
).do( lambda doc:
  //Calculate the squared averages
  return doc.merge({

  #Estimated beta for temperature
    "beta_hat":
     #Numerator
     ((doc("n") * doc("sum_products_temp_humidity")) - (doc("sum_humidity") * doc("sum_temp"))) /
     #Denominator
     (doc("n") * doc("sumsq_temp")) - (doc("sum_temp") * doc("sum_temp"))
    ,

    #Standard Deviation calculations
    "sd_humidity": obs["humidity"].map( lambda d:
        return (d - doc("avg_humidity") * d - doc("avg_humidity"))
    ).avg().do(r.js('(function(x) { return Math.sqrt(x); })')),

    "sd_temperature": obs["temp"].map( lambda d:
        return (d - doc("avg_temp") * d - doc("avg_temp"))
    ).avg().do(r.js('(function(x) { return Math.sqrt(x); })')),

    #Correlation calculation
    "temp_humidity_correlation":
      doc("sum_products_temp_humidity") / ((doc("sumsq_temp") * doc("sumsq_humidity")).do(r.js('(function(x) { return Math.sqrt(x); })')))
  })  

#Finally calculate alpha estimator
).do(lambda doc:
  return doc.merge({
    "alpha_hat":
      ((r.expr(1) / doc("n")) * doc("sum_humidity")) - (doc("beta_hat") * (r.expr(1) / doc("n")) * doc("sum_temp"))
  });
).run(conn)