A new system for teaching materials

BeeBIT has always relied on the commitment of interested teachers to prepare the data collected by the eHives for teaching in schools. As a handout, we provide a large number of elaborated lessons for free download. The materials cover all science subjects from mathematics to computer science, physics, chemistry, biology and geography. Whether elementary school or high school, suitable questions and difficulty levels can be filtered in our database, viewed, downloaded and then edited if necessary. (All teaching materials are offered as docx files to allow teachers to make adjustments based on the curriculum and students' qualifications).

In addition to the freely available worksheets, there is a protected area on our website that contains suggested solutions and materials specifically intended for teachers, which can only be accessed after registration. BeeBIT checks once whether new users are authorized to access these materials based on their professional background. This prevents crafty students from unauthorizedly downloading the solutions to the worksheets. The system for managing and downloading the lessons has not been further developed since 2016.

Fig. 1: Screenshot of the old system. One can see the gray color scheme and the confusing listing of materials without any standard structuring according to available languages or other categories.

The new system

After launching the modernized BeeBIT website in March 2019, the provisional nature of the old system became more and more obvious and the association decided to tackle a modernization. After a long time and several attempts, we can finally announce the launch of the new user interface. You can find it in the navigation menu of the site under the tab 'Didactics'. Besides a more appealing visual design, some new or improved functionalities could be implemented, which will be briefly explained in the following. To conclude this blog post, we would like to give a small insight into the spectrum of available materials. The goal is to motivate readers to discover the materials on their own.

Fig. 2: The new system for browsing, filtering and downloading teaching materials. Note the clearer layout under Summary of all languages offered in a module card including listing of author, subject, topic and target group/level.

Presentation and Navigation

As can be seen in the screenshot above, all materials are now displayed by default in grouping of all available languages. This makes it easy to see whether a suitable translation is available. In addition to the title, further information such as author, subject or tags are listed directly, enabling a quick overview without requiring the explicit opening of individual files. By clicking on property values marked with a black underline, a filter is activated that only displays materials with the same value. Clicking on the colored module card of each material takes you to the detailed view including a brief description and the links to download.

Related materials / Suggestions

In the detailed view, related materials from our database are automatically listed at the bottom of the screen. If the selected lesson suggestion does not quite meet your needs, it is worth taking a look at this listing. Although the suggestions are generated in different categories (author, subject, tags), we have made sure that there are no annoying duplications in the suggestions.

Login system

Access to the teaching staff area is restricted by a login (as in the old system). However, a user session is now valid on the entire BeeBIT homepage (with the exception of the diagram viewer). While the site does not yet effectively take advantage of such an overarching system, future scenarios are conceivable that could require a login to unlock supplemental materials in other areas of the site as well. Teachers who do not yet have an account can easily register. Please note that the login data could not be transferred from the old system. For a short transition period, both systems will be operated in parallel. Nevertheless, we ask all those affected to register for the new system. A renewed check of your access authorization will not be necessary.

A peek into available materials

Representative of the large number of teaching units in our database, three examples will be presented to give a rough overview of the scope, orientation and content of the materials.

Forest or flower honey

The first lesson stems from a physical-chemical area. The students receive an introduction to the methodology for measuring the electrical conductivity of a liquid. The conductivity can be used to distinguish between different types of honey. Quote from the introductory section of the worksheet:

Electrical conductivity is considered to be one of the best parameters for analysing and authenticating honey, and differentiating between honey made from nectar (“blossom” or “flower honey”) and honey made from honeydew (“forest honey”). It is the origin that determines the composition of the honey – and hence its electrical conductivity. According to the guidelines of the International Honey Commission (IHC), the electrical conductivity of honey is defined as that of a 20 % weight in volume (w/v) solution in distilled water at 20 °C, where the 20% refers to honey dry matter (dry matter basis). The result is expressed in microSiemens per centimetre (µS/cm). Honeys with an electrical conductivity of more than 800 µS/cm are labelled honeydew or forest honeys. The electrical conductivity values of mixtures of forest and flower honeys range between 600 and 800 µS/cm. Pure flower honeys show electrical conductivity values between 350 and 600 µS/cm. If the electric conductivity value is below 250 µS/cm, this might indicate that the bees have been artificially fed with sugar and such honey is banned by Departments of Health and Food Safety in many countries.

If the water content of the honey to be tested is known (see related experiment Water in honey), the dry mass required for the experiment can be weighed out. Only with precise weighing one receives a comparable value of the electrical conductivity after dissolving the honey in distilled (!) water. By measuring different types of honey, the students should validate the values claimed in the introductory text. How can an unacceptable dilution of the honey be excluded using the learned method? Is there a correlation between colour and electrical conductivity of the honey?

Bee larvae

In this strongly biologically influenced lesson, the students should first learn about the spatial organization of a honeycomb using the photo below and an explanatory text. Then, the temperature inside a beehive is investigated. Based on the measurement data of the eHives, students should compare the inside and outside temperature. What is noticeable at the time of brood? Can the position of the brood combs be determined from the temperature curves? The author gives the following hint:

The bees have to maintain a constant temperature of 35°C in the hive, so that the larvae can develop into adults.

Abb. 3: Honey combs with different areas.

(Note: For a more in-depth analysis of the eHive data, we refer to posts in this blog. For example, an analysis using the Python3 programming language is presented in detail in the posts of 07/20/2019 and 09/22/2019).

The bee database (part 1/3)

The topic of databases in the subject of computer science is, from experience, a rather dry undertaking for many students. The author of the last material presented here has designed three practical lessons, the first of which is presented as an example.

The data collected by the eHives is stored in an SQL database and can also be displayed via the diagram viewer using SQL queries. Starting from real measured data, basic elements of the syntax are introduced and learned in small practical questions. From the table of the outdoor temperature sensor, for example, the highest and lowest temperature ever measured in the observation period can be queried quite easily.

Parts 2 and 3 of the teaching module deal with increasingly complex questions but at the same time allow the investigation of more interesting correlations. For example, can a correlation be derived between weather and the development of stick weight? What SQL queries can be asked to help answer the question?

(cw) 2021-10-03


Newsletter June 2020

We are happy to announce the publication of our latest newsletter. You can download the PDF by clicking on the teaser below.

(team) 2020-06-08


The eHives' locations and surroundings

The following article gives a brief overview about the eHives' locations and surroundings in tabular form. The position of the weather station and its appendant wind sensor are described in a few words. For both, the first value in the table denotes the mounting height. It was tried to uniformely position the weather station and its wind sensor two respectively five meters above ground. However, due to the local situation this was not possible everywhere.

Fig. 1: The eHives' locations. Honey-coloured points mark locations of currently active eHives, dark grey points mark previous locations.

Fig. 2: Timeline for the eHives' activity periods.

DEU-DHG-1 & DEU-FKG-1

Activity period DEU-DHG-1: 12.05.2016 -
DEU-FKG-1: 27.06.2016 -
eHive Version 0.2.2, since 21.08.2018: 0.2.1
Coordinates 49.79, 9.92
Surroundings lawn, partially covered by trees behind the hives, located in close distance to a small garden shelter
Weather station 2 m, mounted in close distance to several trees
Wind sensor 5 m, about the height of the trees' crowns
Covered by a roof? ✗ no

AUT-GSC-1

Activity period 23.04.2016 -
eHive Version 0.2.2, since 2019: 0.2.1
Coordinates 48.20, 16.39
Surroundings roof of a three-storey house, nearby building behind the eHive is one level higher
Weather station 1 m, mounted on the eHive's roof
Wind sensor 1 m, mounted on the eHive's roof
Covered by a roof? ✓ yes

AUT-WIS-1

Activity period 18.05.2019 -
eHive Version 0.2.1
Coordinates 48.24, 16.415
Surroundings green area, hive is covered by trees, located in close distance to a building
Weather station 4 m, mounted on the edge of a balcony, a bit lower than the rooftop
Wind sensor 5 m, mounted on the roof gable
Covered by a roof? ✗ no

AUT-BIE-1

Activity period 25.05.2016 -
eHive Version 0.2.2
Coordinates 48.26, 16.48
Surroundings below a roof overhang directly in front of a building
Weather station 1 m, in front of small bushes
Wind sensor 2 m
Covered by a roof? ✓ yes

ITA-FEM-1 & ITA-FEM-3

ITA-FEM-3 currently under the name ITA-FEM-2

Activity period ITA-FEM-1: 18.10.2017 - 13.10.2018 under the name AUT-WIS-1, 09.10.2019 -
ITA-FEM-3: 09.10.2019 -
eHive Version ITA-FEM-1: 0.2.2, since 09.10.2019: 0.2.1
ITA-FEM-3: 0.2.1
Coordinates 46.07, 11.23
Surroundings terrace, distance to a nearby building approx. two meters
Weather station 1 m above the terrace's height level, at the edge of the terrace 5m away from the building, 4 m above the ground level of the lawn in front of the terrace
Wind sensor 7 m above the terrace, a bit higher than the nearby building
Covered by a roof? ✗ no

ITA-FEM-2

Activity period 02.01.2017 - 14.10.2018
eHive Version 0.2.2
Coordinates 46.36, 10.92
Surroundings lawn, located at a hillside
Weather station 2 m
Wind sensor 4 m
Covered by a roof? ✗ no

POL-LOK-1

Activity period 03.02.2017 - 11.07.2018 under the name POL-LOK-2, 18.11.2019 -
eHive Version 0.2.2, since 18.11.2019: 0.2.1
Coordinates 50.57, 21.67
Surroundings inside a garden, in front of a fence, surrounded by bushes
Weather station 2 m
Wind sensor 4 m
Covered by a roof? ✗ no

DEU-EUR-1

currently under the name DEU-BGT-1

Activity period -
eHive Version 0.2.2
Coordinates 49.67, 10.04
Surroundings inside a small open shelter that is limited by a building and a wall in two opposite directions
Weather station 2 m
Wind sensor 3 m
Covered by a roof? ✓ yes

ITA-LFV-1

Activity period 29.12.2016 - 12.10.2018
eHive Version 0.2.2
Coordinates 46.38, 11.24
Surroundings inside a bee house
Weather station 4 m, directly above the rooftop
Wind sensor 5 m, approx. one meter above the rooftop
Covered by a roof? ✓ yes

DEU-MNG-1

Activity period 15.02.2018 -
eHive Version 1.1
Coordinates 51.19, 6.44
Surroundings roof terrace at the height of the second floor, the terrace is enclosed by the building which is two floors higher in two opposing directions and one floor higher in the two remaining directions
Weather station 2 m
Wind sensor 4 m
Covered by a roof? ✓ yes

DEU-OEG-1

Activity period 06.09.2017 -
eHive Version 1.1
Coordinates 53.10, 8.92
Surroundings roof terrace of a two-storey building, the roof enclosing the terrace is approx. one level higher
Weather station 2 m
Wind sensor 5 m
Covered by a roof? ✓ yes

DEU-FDG-1

Activity period 03.04.2018 - 13.02.2019
eHive Version 0.2.1
Coordinates 49.97, 9.13
Surroundings lawn, located near some buildings
Weather station 2 m
Wind sensor 5 m, located near a three-storey building
Covered by a roof? ✗ no

DEU-BBT-1

currently under the name DEU-FDG-1

Activity period 07.05.2019 - 08.02.2020
eHive Version 0.2.1
Coordinates 49.79, 9.88
Surroundings garden, a few meters in front of a house, covered by a tree
Weather station 2 m, in three meters distance to the house
Wind sensor 3 m, in two meters distance to the house
Covered by a roof? ✗ no

DEU-LPG-1

Activity period 08.04.2019 -
eHive Version 1.2
Coordinates approx. 48.14, 11.59
Surroundings -
Weather station 2 m
Wind sensor 5 m, near a three-storey building
Covered by a roof? ✓ yes

(jg) 2020-03-15


2019 visualized

Happy New Year!

The BeeBIT team wishes you a happy new year 2020!

Since the start of the project in 2013 a lot of work has been put in development and operation of the eHives. In 2015, the first eHives were shipped and the association »BeeBIT e.V.« was founded. Our database is growing daily ever since. As the decade ends, we had a brief retrospective of 2019 in mind to thank the people and institutions that are working with our/their eHives. Enjoy the following text and make sure to check out the supplementary materials!

2019 visualized: AUT-BIE-1

The aim of this blog post is to show a huge amount of data (often more than 60 MB per eHive) in a figure that is small in size and reasonably easy to understand. We picked eHive AUT-BIE-1 as an example because this hive collected data over the whole year 2019 without interruption. However, other eHives with enough data were not ignored and a figure similar to the one shown beneath was created and can be downloaded from our website, c.f. supplementary materials at the bottom of this blog post.

Don't worry, the following figure is explained in detail in the text underneath.

Figure: Full data from 2019 of eHive AUT-BIE-1. For visualization the daily mean of all datasets was computed. More details are explained in the text underneath.

Outside Temperature & Solar Irradiation

The data from the temperature sensor of the weather station (as well as all other data shown) was downloaded from the website's diagram viewer and averaged over 24 hour long intervals. The calculated daily mean is plotted as a line graph where the lines connect daily mean values (365 points) and are colour-coded according to the mean value of the two connected points. The colourbar is shown on the right. On the x-axis the beginning of each month is labeled for better orientation. Dates were calculated using local summer time (that is UTC+2 for Austria where AUT-BIE-1 is located).

On a second y-axis the solar irradiation is visualized in a bar chart. Please note that since we calculated daily mean values, the irradiation during daytime (especially at noon) may be considerably higher.

Inside Temperature

The temperatures inside the eHive are visualized as one block consisting of 365x6 colour-coded tiles (365 days and 6 sensors). The same colourbar as for the outside temperature is used. This approach makes it easy to track the colony's position inside the hive just by looking at the colour of the tiles: If we take a look at the months March and October, we observe that the colony moved from between sensors 3 and 4 in March to sensors 2 and 3 in October.

Temperature Difference to Target

The breeding temperature of a bee colony is circa 35 °C. We can use this information to track when and at which position inside the hive the colony might breed. Therefore, we visualize the deviation from this target temperature. Of course, we have to use a different colorbar for this task. If the deviation is small for an extended period of time, it is very likely that we found the breeding combs. In our example breeding begins around the midst of March and ends mid-August to September.

Hive Weight, Weight Change Rate & Rain Rate

In the last subfigure, the hive weight is shown in a line graph. The 365 mean values are connected by colour-coded lines. The colour corresponds to the weight change rate between two connected points (and is therefore a measure for the line's inclination). Again, the colourbar can be seen on the right. In the data, some jumps with a change rate of more than 1 kg/day can be seen. Most likely, this is a result of the beekeeper's work, e.g. adding or removing a honey shroud. A sudden drop in weight (especially during summer) might also signalize a swarming event.

On a second y-axis the rain rate is visualized in a bar chart. Please note that mm/day is equal to litre/(m² day), using 1 litre = 1 dm³. By taking a look at the beginning of June (around the 6th or 7th), one can see how extensive rainfall leads to a decrease in the weight change rate.

Supplementary Materials

Data from the following eHives was visualized:
DEU-DHG-1, DEU-FKG-1, AUT-WIS-1, AUT-BIE-1, DEU-MNG-1, DEU-OEG-1, DEU-FDG-1, DEU-LPG-1

(cw) 2020-01-01


Advanced data analysis

Based on an introductory blog article on how to analyse raw data, we now want to examine longterm trends. In this article, we want to identity the environmental factors which contribute most to the weight development inside a beehive. In order to do this, we will use the data by the eHive at the AGES (AUT-BIE-1), since there have been neither technical issues nor human interventions over the course of this summer. We are interested most in the timeframe from end of May to end of July, since this is the time when bees collect most nectar.

For Context: The »bee year« starts and ends with summer solstice, i.e. the calendar beginning of summer on June 21. Until then the bee colony grows, i.e. a lot of bees hatch. After, the queen reduces oviposition and the colony prepares for winter. Additionally, the flowering of most crops is over, and it becomes more difficult for the bees to find food.

Disclaimer: In a highly complex biological system such as a bee colony a lot of different processes can influence the weight development. This article concentrates on only one colony, and makes some exemplary observations based on a few select weather values. This of course doesn’t prove that a statement is valid for most or all bee colonies. Furthermore, found correlations don’t imply causation. Instead, the article is supposed to demonstrate what can be examined from eHive data.

1) Data Download

First, we will download the data from the diagram view on the BeeBIT homepage. As working hypothesis, we assume that the weight development correlates with solar irradiation and exterior temperature.

Why especially these environmental factors? We expect the hive weight to increase when the bees collect most nectar and pollen. This implies to conditions that have to be fulfilled:

  • The bees must be able to fly out, so it should be appropriately warm, and there should be neither storm nor rain.
  • Plants must provide nectar. This depends on many factors, but basically it should be neither too wet (rain washes the nectar away) nor too hot (nectar dries out).

Therefore, we included the data Hive »weight«, »Exterior Temperature«, »Interior Temperature 3«, »Solar irradiation« and »Exterior humidity« in the dataset.

Fig. 1: Data displayed by the diagram viewer on BeeBIT’s website.

In Fig. 1, we only detect a few general trends: the hive weight increases steadily until end of June and then slowly decreases in July, also the interior temperature is nearly constant at about 35°C. With NumPy we can calculate daily averages of the data. For the weight, we don’t need the average but the difference of two values at midnight, c.f. Fig. 2.

Fig. 2: Daily mean values of the data. The differences in the hive’s weight are calculated at midnight.

2) Influences on the hive weight, June

Still too chaotic? We will simplify the data even further in a minute. We spot in the first month (before summer solstice), that a high solar irradiation (green curve) correlates with a high increase in hive weight (blue curve), and that the weight increase is rather low (or even negative) on days with low solar irradiation. Later, this correlation appears to become weaker.

Apparently, something changes within the colony, and this possibly has something to do with summer solstice. We now plot all data points of the first 40 days (up to 2019/07/08) in a scatter plot, c.f. Fig. 3. The horizontal position of a day is determined by the weight difference, the vertical position by the solar irradiation. The color indicates the exterior temperature.

Fig. 3: Correlation between solar irradiation and weight difference for the first 40 days of the observation period. The outside temperature is shown color-coded.

The presumed correlation apparently really exists! On days with high solar irradiation, the hive weight increases more than on days with low solar irradiation. We assume that larvae don’t grow faster or slower depending on solar irradiation, so the observed difference probably comes from collected nectar. There are, however, a few outliers, which we will discuss later: days with high solar irradiation where the hive weight decreases (top left corner) and days with low solar irradiation where the hive weight increases (bottom right). These outliers are labeled with their day, so we can find them later.

Also, the weight change appears to be relatively independent from the exterior temperature in June.

3) Influences on the hive weight, July

This changes in July, therefore we exchange the horizontal axis. The color of the points now correlates with solar irradiation, c.f. Fig. 4. The plot contains all days starting from day 25 (June 23).

Fig. 4: Correlation between outside temperature and weight difference starting at day 25 of the observation period. The solar irradiation is shown color-coded.

Apparently, the exterior temperature is the governing factor for weight development after summer solstice. This can be well explained: At the beginning, we saw that the interior temperature remains approximately constant. Consequently, the bees have to actively heat on most days, i.e. consume food – and they do so the most on coldest days. They also consumed food in June, but during days with high solar irradiation they were able to collect more food than they consumed. These two factors appear to govern the weight change. From literature we know that most bees hatch before summer solstice, which contributes to the high difference of the absolute values of weight change in June and July.

4) Outlier

We notice that there are outliers in both scatter plots. Ideally, we can find an explanation for these using the BeeBIT data. As example, we concentrate on the outliers in June.

Fig. 5: Outlier. Red points and black crosses mark days with notably high respectively low weight change in respect to days with similar solar irradiation.

We still have some weather factors which have not been analyzed yet: Precipitation and air pressure. When we mark the outliers of the first scatter plot, we discover that they often appear near heavy rain, often the day before or after, c.f. Fig. 5. The red points are days where more nectar was collected than on days with comparable solar irradiation, the black crosses are days where less than usual is collected.

For the two black points, we can find a simple explanation: On the day after rain, the flowers are still wet, the bees have difficulty reaching the nectar. Therefore, they are collecting less, even if the sun is shining.

The red points appear either on the days before rain or one or two days after rain. Since the bees appear to target a weight increase of 0.5 kg/day, the work more on days after rain (compensation of lost time). Furthermore, plants might provide more nectar during these days. Also they are more productive on days before rain (which can be anticipated from lower air pressure), see days 7 and 23. The bees appear to factor future weather changes into their schedule, and know that they have to make up leeway after losses during rainy days. I was unable to find this observation in the literature, and it should be evaluated more thoroughly. Unfortunately, there currently are very few other datasets, where there have been neither human interventions nor technical problems nor swarm events during an entire summer (i.e. no sudden jumps in weight). As soon as we’ve developed a working bee counter, this hypothesis could be verified or falsified.

5) Summary

  • In the main blooming period in June, the weight increase correlates with solar irradiation and less with temperature.
  • In July the weight difference correlates with temperature and less with solar irradiation, since the bees must heat the hive.
  • Outliers appear on days before or after rainfall. The bees seem to anticipate the rain and compensate bad days on the following days, even if the weather is not ideal for nectar collection.

Some factors of influence have not been discussed due to insufficient data, i.e.:

  • Plants need water to produce nectar.
  • We are unable to decouple weight increases due to reproduction and weight increases due to nectar collection.

Supplementary Materials

(jh) 2019-09-22