The dataset files

The dataset page offers the data as an Excel workbook and as a CSV table. Both contain the same table, built for analysis: one row per electrospinning run, one column per value, every number in a fixed unit.

What's in the Excel workbook

Sheet	Contents
README	Version, date, number of runs and publications, license, how to cite, and how to read the data.
Runs	The data. The header row and record_id stay in view while you scroll, and every column has a filter. Hover over a column header to see its description and unit.
Publications	One row per publication: DOI, link, title, abstract, device and number of runs.
Dictionary	Every column of Runs: group, unit, type and description, and the bulk-upload column it corresponds to.

The CSV is the Runs sheet. It is UTF-8 encoded with a byte-order mark, so Excel shows µ, ° and accents correctly, and it is comma separated.

How to read the data

• Units are in the column names. voltage_kV is always in kV, flow_rate_mL_h in mL/h, fiber_diameter_nm in nm. Values entered in another unit (V, µL/min, µm, mm, K, GPa, …) were converted.
• A run without a recorded unit is in the contribution form's default unit, which is the one in the column name.
• Empty means "not reported". There are no placeholders such as - or 0.
• Several values in one cell (morphology, instabilities) are separated by ; .
• Polymers, solvents and additives are listed together (polymers: PCL + GELATIN) and one per numbered column:
  • polymer_1, polymer_1_mw_kDa, polymer_1_share_wt_pct, …
  • solvent_1, solvent_1_share_vv_pct, …
  • additive_1, additive_1_amount, additive_1_amount_unit (wt% or mg/mL), additive_1_amount_basis (what the amount is relative to: polymer, solution or solvent), …
• Copolymers have polymer_N_monomers and polymer_N_composition, e.g. LA:GA and 75:25, in polymer_N_composition_unit. polymer_type is homopolymer, copolymer, blend or core/shell.
• Mixed and core/shell solutions (see Solution preparation):
  • solution_preparation says which (single, mixed, core/shell), and solution_description summarises the solutions.
  • The full structure is in the JSON download and the API.
  • Coaxial runs have flow_rate_core_mL_h and flow_rate_shell_mL_h, and flow_rate_mL_h is their total.
• Fiber diameter: fiber_diameter_qualifier is exact, approximate or range. For a range, fiber_diameter_nm is the midpoint and fiber_diameter_min_nm / fiber_diameter_max_nm are the bounds.
• Concentration keeps its unit (concentration, concentration_unit, mostly wt%). wt% and mg/mL cannot be converted into each other without the solution's density.
• unit_notes lists any value whose unit could not be converted; such values are left empty instead of being guessed.
• completeness_score (0–100) shows how much of the important information a run has.
• record_id identifies a run and stays the same across dataset versions.

Column groups

Group	Columns
Run	record_id, completeness_score
Publication	doi, publication_url, publication_title, data_origin, extraction_method, source_location
Polymers / Additives / Solvents	polymer_type, polymers, polymer_N_…, additives, additive_N_…, solvents, solvent_N_…
Solution	solution_preparation, solution_description, concentration, concentration_unit, viscosity_cP, surface_tension_mN_m, conductivity_mS_cm, evaporation_rate_mg_cm2_h, ph
Needle	needle_type, needle_gauge_G, needle_diameter_mm, needle_inner_gauge_G, needle_inner_diameter_mm, needle_middle_gauge_G, needle_middle_diameter_mm, needle_outer_gauge_G, needle_outer_diameter_mm, needle_material. A needle size is either a gauge or an inner diameter in mm, as reported; gauges are not converted to mm.
Collector & device	collector_type, collector_material, collector_…_cm / _rpm / _mm, device_manufacturer, device_model, custom_device
Process	voltage_kV, flow_rate_mL_h (+ _core_ / _middle_ / _shell_), tip_collector_distance_cm, spinning_duration_min, temperature_C, relative_humidity_pct, formation_stable, instabilities
Fibers	fiber_diameter_nm, fiber_diameter_sd_nm, fiber_diameter_qualifier, fiber_diameter_min_nm, fiber_diameter_max_nm, diameter_method, diameter_n, morphology, product_weight_mg
Properties	tensile_strength_MPa, youngs_modulus_MPa, elongation_at_break_pct, fracture_behaviour, specific_surface_area_m2_g, porosity_pct, permeability_nm2, water_contact_angle_deg, electrical_conductivity_S_m, thermal_conductivity_W_mK, application fields
Quality	unit_notes

The Dictionary sheet is the complete, authoritative list. The numbered columns start at three per component and grow if a run has more.

Loading the data

import pandas as pd

runs = pd.read_csv("electrospinning_dataset.csv")
# or: pd.read_excel("electrospinning_dataset.xlsx", sheet_name="Runs")
pan = runs[runs["polymers"] == "PAN"]
pan.plot.scatter(x="voltage_kV", y="fiber_diameter_nm")

runs <- read.csv("electrospinning_dataset.csv", fileEncoding = "UTF-8-BOM")
summary(runs$fiber_diameter_nm)

Correcting or extending the data

The downloaded file (Excel or CSV) can be uploaded on the bulk upload page, and the web form's Import from Excel reads it too.

• New runs: the rows are uploaded as new runs. Use this to add your own runs in the same layout.
• Corrections: to correct a run that is already published, edit it in My Submissions or contact us.
• Mixed and core/shell runs are re-uploaded as one solution; describe them on the Solutions sheet to keep their structure.

The image dataset

The Image dataset (.zip) download (and format=zip on the export endpoints) holds the fiber images of the runs, with their metadata:

README.md     version, license, citation, counts, how to load it
images.csv    one row per image
runs.csv      the runs that have images, in the same columns as the main dataset
images/{record_id}/{image_id}_{type}.{ext}

• Files: images are grouped by run. The extension is the file's real format (png, jpg, tif, ...), detected from the file itself.
• images.csv: image_id, file (path in the archive), record_id, image_type (SEM, FESEM, Cross-sectional SEM, Microscopy, Simple Image), file_format, width_px, height_px, file_size_bytes, sha256, scale_bar_nm, magnification_x, staining, description, other_metadata, followed by key values of the run (doi, polymers, solvents, concentration, voltage_kV, flow_rate_mL_h, tip_collector_distance_cm, fiber_diameter_nm, morphology, ...). For everything else, join with runs.csv on record_id.
• A version's zip is made from the same runs as its tables, so its values match the Excel and CSV of that version.

import pandas as pd
from PIL import Image

images = pd.read_csv("images.csv")
sem = images[images.image_type == "SEM"].dropna(subset=["fiber_diameter_nm"])
for row in sem.itertuples():
    img = Image.open(row.file)
    target = row.fiber_diameter_nm

Dataset versions published before October 2026 keep the layout they were released with.