Open Science

In the SteeleLab, we are committed to open science. On this page, you will find a list our open science efforts, including information about open access, open data, our open source software projects, work related to open hardware, and some of my efforts at engaging dialogue with scientists about open science.

Open Access

We strive to have all of our publications available open-access (“green”) on the arxiv server (see publications page).

Open Data

In the group, we have started by working with a few basic levels of open data:

  • Level 0: Publication of the processed data files as plotted in the figures in the paper.
  • Level 1: Publication of the raw data, as recorded by our computers, along with software and processing scripts that derive the plotted data from the raw data These definitions are based on a QN department policy that I developed with Anton Akhmerov.

In level 1, we may also include the measurements scripts and that controlled the equipment when the data was measured. In the group, our scripts typically use the open source library STLab, making it possible to fully reproduce the measurement procedure performed.

We are also looking towards the future beyond open data. For this, we are working on developing practical, pragmatic ways of more thoroughly documenting the research we do using the tools that modern ICT can offer.

Examples of publications including Level 1 Open Data:

  • A ballistic graphene superconducting microwave circuit
    Felix E. Schmidt, Mark D. Jenkins, Kenji Watanabe, Takashi Taniguchi, Gary A. Steele
    Nature Communications 9, 4069 (2018) arxiv

  • Observation and stabilization of photonic Fock states in a hot radio-frequency resonator
    Mario F. Gely, Marios Kounalakis, Christian Dickel, Jacob Dalle, Rémy Vatré, Brian Baker, Mark D. Jenkins, Gary A. Steele
    Science 363, 1072 (2019) arxiv

Examples of publications including Level 0 Open Data:

  • Bimodal Phase Diagram of the Superfluid Density in LaAlO3/SrTiO3 Revealed by an Interfacial Waveguide Resonator
    Nicola Manca, Daniel Bothner, Ana M. R. V. L. Monteiro, Dejan Davidovikj, Yildiz G. Sağlam, Mark Jenkins, Marc Gabay, Gary Steele, Andrea D. Caviglia
    Phys. Rev. Lett. 122, 036801 (2019) arxiv

  • Large spin-orbit coupling in carbon nanotubes
    G.A. Steele, F. Pei, E.A. Laird, J.M. Jol, H.B. Meerwaldt & L.P. Kouwenhoven
    Nature Communications 4:1573 (2013) arxiv

Data management and infrastructure

In the SteeleLab, as experimentalists, we take data extremely seriously: for experimental physics, data is our life blood. Ever single point of data we take off our instruments is sytematically archived and made accessible, along with the measurement scripts that created the data. We also make sure we rigorously keep records of all the code we use to process our data, and make all of this publicly available following the open data “level” systemisation we describe above. You can find our official SteeleLab DMP here:

SteeleLab Data Management Policy

Making lofty statement is one thing: lots of high up people can do this without any regard to the implications for the people “on the workfloor”. That is why for us what is equally or perhaps even more important is the deployment of a practical and pragmatic infrastructure to allow this to happy without overhead and headaches for the people doing the research (data management should reduce work load, not increase it).

How do we do all of this in our group? First, I (personally) have hacked together a robust infrastructure for ensuring that every piece of data taken is automatically transferred to a central location, is backed up regularly, and is accessible to everyone in the group through a uniform, cloud-deployed computational environment (jupyterhub):

SteeleLab ICT Infrastructure

Second, we ensure that our measurement scripts implement robust time-stamped, folder-based data saving policies. Crucial is libraries that facilitate this, such as the recent “DataFolder” functionality I implemented in our data explorer library that now decouples this from the specific measurement libraries to achieve uniform and robust data storage:

Using DataFolder

One simple key aspect of this library is automatic duplication of the measurement script in the folder where the data is stored so that one can have certainty of how the data was generated.

Open Source Software

All of our code is released under open source licences such as the GPL.

  • Spyview A data visualization / anlaysis software package that Gary wrote during his PhD.
  • STLab A python / VISA based measurement library coded from the ground up our former postdoc Mark Jenkins.
  • STCad A python library for mask design of our chips, including resonators, qubits and DC devices. Note that the designs produced by the example code in the library are licensed under the CC-BY-SA licence.
  • QuCAT A Quantum Circuit Analysis Tool developed in the group by Mario Gely for drawing and performing quantisation analysis of circuits.
  • Logbook generator A library for automatically generating a basic electronic lab notebook (ELN) for documenting data. Supports data files with user-defined plotting routines, and also image logbooks for generating notebooks based on hierarchies of folder containing image.

Open Hardware

Below, you can find a full bill of materials and details of contruction of our 2D material stamping setup. We are currently considering making more of our hardware projects fully open.

Open Science Dialogue

I think it’s important to stimulate dialogue with scientists about Open Science. Here are some of my efforts in this direction: