Today we're going to discuss how machine learning can be used to group and label information even if those labels don't exist. We'll explore two types of clustering used in Unsupervised Machine Learning: k-means and Hierarchical clustering, and show how they can be used in many ways - from book suggestions and medical interventions, to giving people better deals on pizza!

Following lecture presentation of classification schemes and environments of formation for non-siliciclastic rocks, students perform a sequence of two activities (one in the lab, one in the field) designed to guide them from theoretical understanding to practical application. In the lab component, students are presented with twenty hand samples and twenty rock names and/or descriptions. Their task is to examine the samples, match the descriptions with the proper rocks, and propose a likely environment of formation for each sample. Included among the samples in the lab component are carbonates and cherts similar to those encountered in the field component, which examines a local outcrop exposing four successive carbonate formations in the Middle Paleozoic Helderberg Group. In the field component, students are first asked to differentiate between successive formations using lithology, sedimentary structures, and fossil assemblages. They then determine the probable sedimentary environment for each formation, and interpret the facies succession preserved in the outcrop to reconstruct a small portion of local geologic history. In both field and lab components, students are encouraged to work in small groups to develop their initial responses without instructor input. This arrangement ultimately improves both student understanding of the material and confidence in their own interpretations.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

Magmatic differentiation is an important concept in the geology curriculum. Students are generally introduced to magmatic differentiation in the introductory course, whereas the details are typically developed in mineralogy and petrology courses. In particular, students often struggle to understand the processes of fractional crystallization.

Fractional crystallization by gravity settling can be illustrated using a model magma chamber consisting of M&M'sÂ. In this model, each major cation (e.g., Si, Ti, Al, Fe, Mg, Ca, Na, K) is represented by a different color M&MÂ; other kinds of differently colored or shaped pieces could also be used. Appropriate numbers of each color M&M are combined to approximate the cation proportions of a basaltic magma; this is the "parental magma". The M&M's are then placed in a group on a tabletop to form a magma chamber. Students then fractionate the magma in ten crystallization steps. In each step the M&M's are moved to the bottom of the magma chamber forming a series of cumulus layers; the M&M's are removed in proportions that are identical to those of the stoichiometric proportions of cations in the crystallizing minerals (e.g., olivine, pyroxene, feldspars, quartz, magnetite, ilmenite). Students observe the changing cation composition (proportions of colors of M&M'sÂ) in the cumulus layers and in the magma chamber and graph the results using spreadsheet software. Students classify the cumulates and resulting liquid after each crystallization step, and they compare the model system with natural magmatic systems (e.g., absence of important fractionating phases, volatiles). Students who have completed this exercise demonstrate increased understanding of fractionation processes exhibit greater familiarity with mineral stoichiometry, classification, solid-solution in minerals, element behavior (e.g., incompatibility), and chemical variation diagrams.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

Students conduct autocorrelation and cross-correlation analyses on river discharge and climate indices to test the hypothesis that coastal streams draining mountainous terrain are strong indicators of climatic phenomena. Students must load the data, conduct analyses, and plot the results by writing an efficient MATLAB script, and must "publish" their code into a well-organized, well-commented, .pdf document (or .html on a website).

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

The student compiles oil field well data, including spatial locations and digital well logs, for analysis of subsurface, oil reservoir stratigraphy and lithology using computer spatial technologies software, Geoplus Petra. The students create well log cross sections, make lithologic picks, construct structure and isopach maps, and evaluate lithologic properties, including gross reservoir quality from petrophysical logs. These data are used to interpret depositional environment of the subject formation and make predictions for well bore perforations for oil production. The key value of the exercise is an introduction to the use of computer software to analyze geological data, guided by sedimentologic and stratigraphic insights, and make predictions for resource exploitation.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

How can we use data to improve ecological forecasts?
To be useful for management, ecological forecasts need to be both accurate enough for managers to be able to rely on them for decision-making andinclude a representation of forecast uncertainty, so managers can properly interpret the probability of future events. To improve forecast accuracy, we can update forecasts with observational data once they become available, a process known asdata assimilation. Recent improvements in environmental sensor technology and an increase in the number of sensors deployed in ecosystems have resulted in an increase in the availability of data for assimilation to help develop and improve forecasts for natural resource management. In this module, students will develop an ecosystem model of primary productivity, use the model to generate forecasts, and then explore how assimilating different types of data at different temporal frequencies (e.g., daily, weekly) affects forecast accuracy. Finally, students will assimilate different types of data into forecasts and examine how data assimilation affects water resource management decisions.

Because of increased variability in populations, communities, and ecosystems due to land use and climate change, there is a pressing need to know the future state of ecological systems across space and time. Ecological forecasting is an emerging approach which provides an estimate of the future state of an ecological system with uncertainty, allowing society to preemptively prepare for fluctuations in important ecosystem services. However, forecasts must be effectively designed and communicated to those who need them to realize their potential for protecting natural resources.
In this module, students will explore real ecological forecast visualizations, identify ways to represent uncertainty, make management decisions using forecast visualizations and learn decision support techniques. Lastly, students will then customize a forecast visualization for a specific forecast user's decision needs.
The overarching goal of this module is for students to understand how forecasts are connected to decision-making of forecast users, or the managers, policy-makers, and other members of society who use forecasts to inform decision-making.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

This exercise is part of a sequence of exercises to help students understand single and multiple aquifer tests. In class, students will receive data from pumping tests at University of Minnesota's hydrogeology field site. Instead of using commercial software, students will create interactive modeling tools to facilitate curve matching and to collaborate on understanding aquifer tests at different scales using different methods. The approaches used in this exercise can be extended to develop models to compare single and multiple aquifer tests using different approaches.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

This computer-based assignment forces students to compare and contrast integral and differential forms of the conservation of mass equation, as well as analytical and numerical approaches to solution. Students are given a text description of a simple environmental problem (a conservative tracer diffusing in a one-dimensional system with no-flux boundaries) and are then required to first write equations that describe the system and then implement these equations in an Excel spreadsheet or Matlab m-file. Students then use their spreadsheets/m-files to compare different solution methods and must communicate these results in short text answers.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

Students will predict how many amusement parks are in their state. They will then analyze census data on the numbers of amusement parks in all 50 states in 2016. (Data in this activity do not include the District of Columbia or Puerto Rico.) Then students will write numbers as fractions and create a visual model of the data.

This activity uses field measurements and GIS to estimate the volume of water in the form of snow in a field site.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

This lab uses Google Earth to measure the rate of seacliff retreat. It touches upon coastal processes, natural hazards, and coastal management issues. The central focus of the lab is in the Monterey Bay area.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

Students are introduced to concepts of sampling distributions, p-values, and hypothesis testing. Using both simulated and real data for methylmercury level in fish populations, students will determine whether observations fall within government safety guidelines for safe consumption.

Students will explore an interactive data visualization of state-by-state population growth as measured by the decennial censuses of 1790 through 2010. Students will also analyze and make inferences about the causes of more recent shifts in U.S. population.

Using Lab Measurements to determine the power output of a solar module and the economic feasibility of photovoltaic panels

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

Students are asked to form a persuasive argument using landmarks proving which pair of sneakers they should purchase.

In this adaptation students in online classrooms will use linear regression at home and in synchronous online time to analyze real data on vector-borne diseases and explore how environmental factors affect disease transmission.

In this activity students will use linear regression to analyze real data on vector-borne diseases and explore how environmental factors such as climate change or population density influence the transmission of these diseases.

Students are challenged to a Sherlock Holmes-style mystery in which they construct their own decay curves of melting ice to determine time-zero.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)

The purpose of this exercise is to guide students through the initial stages of a hypothetical water supply complaint investigation for 3 scenarios where methane has been detected in surface water in PA. This activity uses data accessible through the CUAHSI database (Shale Network (2015)) at three stream locations, and introduces three different potential sources of the observed in-stream methane. The data were collected by personnel from Penn State, the U.S. Geological Survey, and volunteers.

(Note: this resource was added to OER Commons as part of a batch upload of over 2,200 records. If you notice an issue with the quality of the metadata, please let us know by using the 'report' button and we will flag it for consideration.)